Source Research Lab / Trading Systems Research Lane Simulation-First No Signals

Trading Systems Researcher
Quant Benchmark Operator

Help test, benchmark, document, and improve Source’s premium data workflows, market data pipelines, backtesting systems, simulations, broker/exchange maps, risk dashboards, AI market research agents, telemetry systems, local inference labs, and proprietary market research infrastructure.

△ Apply for Trading Systems Research Review View Research Programs → Back to Research Lab
Trading Systems Research
Quant Research
Benchmark Operators
Premium Data Workflows
Backtesting Integrity
Simulation-First

No Signals. No Guarantees. Systems Research Only.

Trading involves substantial risk. AI does not remove risk. Machine learning does not guarantee performance. Backtesting does not guarantee live results. Simulation does not guarantee execution quality. Premium data does not guarantee profitable decisions. No strategy is guaranteed to work.

This lane is educational, research-focused, benchmark-focused, simulation-focused, paper-trading-first, and systems-focused. It is designed to evaluate market-systems literacy, research discipline, infrastructure quality, dashboard utility, data integrity, risk-control thinking, and operator judgment. It is not investment advice. It is not a trading signals service. It is not a guaranteed-profit program. It is not a funding guarantee. It is not a live-trading promise.

This Lane Does NOT Provide:

  • • Investment advice
  • • Trading signals
  • • Guaranteed profits
  • • Guaranteed returns
  • • Guaranteed pips
  • • Guaranteed funding
  • • Account growth
  • • Payout guarantees
  • • Prop-firm passing
  • • Risk-free trading
  • • Managing assets
  • • Broker referrals
Core Principle
"The serious path is research first, data quality first, simulation first, paper trading first, risk control always, and documentation everywhere."
Lane Definition

Trading Systems
Research Lane.

The Trading Systems Researcher lane is the applied market systems research and benchmark lane inside Source Research Lab. Its job is to evaluate whether Source’s market systems are accurate, useful, reliable, risk-aware, auditable, and scalable.

RESEARCH METRICS

This Lane Measures Whether Market Systems Actually Work

We design evaluation frameworks around 10 questions to determine system utility before exposing systems to market risk:

  • • Do premium data feeds improve research quality?
  • • Are data pipelines complete, timestamped, normalized, and auditable?
  • • Are backtests contaminated by bias, leakage, bad assumptions, or overfitting?
  • • Do simulations expose failures before capital is at risk?
  • • Do risk dashboards reveal drawdown, exposure, alerts, and intervention points?
EVALUATION SURFACE

Systems Under Evaluation

Researchers validate the components that form the infrastructure of Source’s quant research stacks:

AI-assisted quant notebooks Premium market data pipelines Ingestion & validation engines Alternative & on-chain datasets Event-driven simulation engines Broker API adapters & WebSocket maps Risk dashboards & kill switch units Trading telemetry stacks

The goal is not to chase signals, but to measure, stress-test, document, and improve the systems that make serious market research possible.

Research vs Training

Not The Training Track.
The Research Layer Behind It.

Do not confuse this with enrollment. The training program is designed to train candidates to operate systems. The research lane is designed to test and evaluate the tools, data, workflows, and dashboards.

SOURCE UNIVERSITY TRADING SYSTEMS PROGRAM

Trading Systems Program

Capability Development Layer
  • Learn market systems and study quant workflows
  • Complete dashboards and data pipeline assignments
  • Understand broker/exchange & simulation workflows
View Trading Systems Program  →
SOURCE RESEARCH LAB RESEARCH LANE

Trading Systems Researcher Lane

Benchmark & Evaluation Layer
  • Validate data pipelines and audit backtest bias
  • Review risk dashboards, map broker connectivity, and check telemetry
  • Evaluate AI market research agents and document failure modes
Active Research Lane Focus

The training program asks: Can this person learn how market systems work? • The research lane asks: Do the market systems themselves work?

"The training program is about becoming capable inside market systems. The research lane is about testing whether the market systems themselves are capable, reliable, auditable, and risk-aware."

Candidate Profile

Who This
Is For.

We require capable operators with analytical discipline. While a PhD is not required, applicants must show structured reasoning and technical familiarity with data pipelines or financial systems.

Quant Researchers
Financial Data Engineers
ML/Systems Developers
Risk Systems Analysts
Simulation Engineers
Benchmark Operators
STRONG FIT INDICATORS
  • • Detail-oriented, data-aware, and disciplined in documenting methodologies.
  • • Capable of writing reproducible backtest audits, checking database constraints, or testing latency logs.
  • • Comfortable testing tedious edge cases, admitting uncertainty, and cataloging systems failures.
WEAK FIT INDICATORS
  • • Seeking trading signals, get-rich shortcuts, prop-firm guarantees, or live-risk accounts without reviews.
  • • Expecting name-dropping, classroom lessons, or premium data feeds without producing research reports.
  • • Unwilling to document methodology, map traces, check assumptions, or follow risk protocols.
Research Surface

What You May
Help Test.

Selected researchers are assigned testing scopes within our validation sandboxes. These tasks focus on checking data integrity, backtesting robustness, risk dashboard configurations, and connectivity logic.

1. Data Feeds

Validate premium, institutional, and developer APIs for completeness and query performance.

2. Data Pipelines

Test ingestion pipelines, symbol normalization, timezone alignment, and missing bar rules.

3. Backtests

Audit notebooks for lookahead bias, survivorship bias, data leakage, and overfitting indicators.

4. Workflows

Validate how ideas move from hypothesis to strategy cards, backtests, and code review.

5. AI Assistance

Test whether LLMs can summarize earnings or review strategy assumptions without hallucinations.

6. Risk Panels

Audit drawdown metrics, leverage limit panels, and manual/automatic kill-switch mechanics.

7. Simulations

Test paper-trading environments for partial fills, order rejections, and latency delays.

8. Connectors

Map broker APIs, exchange endpoints, FIX rules, and streaming WebSocket state transitions.

9. Performance

Review Sharpe, Calmar, rolling drawdowns, slippage ratios, and model regime performance.

10. Telemetry

Test trace logs, metrics, alerts, queue health, connection statuses, and incident logs.

Research Programs

Top 10 Active Trading Systems
Research Programs.

The Trading Systems Researcher lane is organized around ten active research programs. Each program is designed to turn vague market-system capability into measurable research output.

Research Question: Can premium, institutional, developer, alternative, and exchange-level market data feeds improve the quality, reliability, auditability, and usefulness of AI-assisted market research compared to weak retail or generic data sources?
Why It Matters: Market systems are only as good as their data. Bad data can create false signals, misleading backtests, incorrect research conclusions, broken dashboards, bad risk estimates, timestamp errors, execution assumptions that cannot survive live conditions, incorrect regime classification, duplicated records, missing bars, distorted volatility, fake robustness, and false confidence.
Systems / Tools Involved:
  • Institutional Terminals & Platforms: Bloomberg Terminal, Bloomberg B-PIPE, Bloomberg Market Data Feed, LSEG Workspace, Refinitiv Workspace, FactSet Workstation, S&P Capital IQ, ICE Data Services, Morningstar Direct, MSCI, Barra, Axioma, PitchBook, Preqin, Sentieo, Koyfin, YCharts.
  • Exchange & Direct Feeds: Nasdaq TotalView, Nasdaq ITCH, Nasdaq Data Link, NYSE Integrated Feed, NYSE OpenBook, NYSE ArcaBook, CME Market Data Platform, CME DataMine, OPRA options, Eurex, futures depth-of-book, ETF constituents.
  • Developer & Quant APIs: Databento, Polygon.io, Massive market data, dxFeed, Intrinio, Tiingo, IEX Cloud, Financial Modeling Prep, Alpha Vantage, Twelve Data, Finnhub, QuantConnect datasets, historical tick data.
  • Alternative Data: SEC EDGAR filings, earnings call transcripts, analyst estimates, economic calendars, FRED, central bank data, news sentiment (RavenPack), Polymarket feeds, crypto on-chain wallets, stablecoin flows.
  • Crypto & On-Chain Data: Kaiko, Coin Metrics, Glassnode, CryptoCompare, Amberdata, DEX liquidity metrics, funding rates, liquidations, perp order-book structures, exchange reserves, contract logs.
Benchmark Method: Data completeness review, missing bar detection, duplicate record detection, timestamp alignment checks, timezone normalization, symbol normalization review, corporate action adjustments, split/dividend adjustments, stale data, latency comparison, quote/trade consistency checks, order-book depth comparison, OHLCV reconstruction, tick-to-bar aggregation, cross-vendor comparison, backtest sensitivity to data source.
Expected Outputs: Premium data feed comparison report, market data pipeline audit, data-quality scorecard, timestamp alignment report, symbol normalization report, tick/OHLCV validation report, order-book data review, alternative data usefulness memo, vendor comparison matrix.
Access Boundary: Access may include public data, vendor documentation, licensed data where available, sample datasets, sandbox datasets, historical datasets, alternative datasets, simulated datasets, and data pipeline prototypes. Access to Bloomberg Terminal, B-PIPE, LSEG, FactSet, ICE, exchange data, paid APIs, premium datasets, or proprietary data feeds is not guaranteed and may depend on licensing, availability, project scope, qualification, and approval.
Research Question: Can Source create backtesting and simulation workflows that identify misleading assumptions, bias, overfitting, slippage errors, and false robustness before any strategy is considered for paper or live-risk review?
Why It Matters: Backtests are easy to fake accidentally. A backtest can look excellent while being useless because of lookahead bias, survivorship bias, data leakage, poor fee modeling, unrealistic slippage, ignored spreads, parameter overfitting, and ignored drawdown duration.
Systems / Tools Involved: Python, pandas, NumPy, Polars, DuckDB, Jupyter notebooks, vectorized backtesting, event-driven backtesting, vectorbt, backtrader, Zipline, QuantConnect / LEAN, NautilusTrader-style simulation concepts, trade logs, slippage/fee/spread models, out-of-sample data, walk-forward testing, Monte Carlo simulators.
Benchmark Method: Lookahead bias tests, survivorship bias review, data leakage audit, signal-to-execution timing review, slippage and fee sensitivity testing, parameter sweep stability sweeps, out-of-sample tests, walk-forward stability review, Monte Carlo distribution audits, regime split robustness tests.
Expected Outputs: Backtesting integrity report, bias detection checklist, strategy QA checklist, slippage sensitivity report, fee modeling report, parameter robustness report, walk-forward analysis memo, Monte Carlo analysis memo, backtest reproducibility package, paper-to-live readiness checklist.
Research Question: Can Source create structured strategy research workflows that move trade ideas from hypothesis to documented research, backtest, simulation, QA review, and improvement without drifting into hype or unsupported claims?
Why It Matters: Trading ideas are cheap, but research discipline is rare. A serious market research workflow requires hypothesis definition, data selection, assumptions, feature engineering, signal logic, risk model, metrics, limitations, reproducibility, review, and retirement criteria.
Systems / Tools Involved: Strategy cards, research notebooks, Jupyter, Python, data pipelines, market data feeds, feature engineering, signal generation, backtesting frameworks, model evaluation tools, risk dashboards, strategy QA checklists, experiment ledgers, Git repositories, vector databases, AI research assistants.
Benchmark Method: Research completeness scoring, hypothesis clarity scoring, assumption audit, feature explanation review, signal logic review, reproducibility checks, metric completeness verification, failure-case documentation score, risk-control coverage scoring, model card validation, human expert audit.
Expected Outputs: Strategy research workflow template, strategy card template, model card template, research notebook standard, strategy review checklist, research completeness rubric, experiment ledger standard, strategy retirement protocol, strategy QA report, research workflow benchmark report.
Research Question: Can LLMs and AI agents help summarize markets, generate hypotheses, review strategies, detect weak assumptions, analyze research notes, and organize market research without hallucinating or creating false confidence?
Why It Matters: AI can accelerate research, but it can also produce plausible nonsense. Confident language can hide weak assumptions, and statistical hallucinations can degrade strategy evaluation.
Systems / Tools Involved: ChatGPT/OpenAI, Claude, Claude Code, Gemini, DeepSeek, Qwen, Kimi, local open LLMs, Hugging Face models, market research agents, strategy/risk reviewers, trade journal auditors, RAG vector libraries, earnings transcripts, SEC EDGAR filings, macro indices, prompt templates, hallucination audit code.
Benchmark Method: Factuality scoring, source-grounding review, hallucination audit, citation faithfulness checks, weak-assumption detection tests, risk-awareness scoring, overconfidence markers evaluation, news summary accuracy, strategy review usefulness, AI vs human review comparison, local-vs-frontier benchmarks.
Expected Outputs: AI quant research evaluation report, LLM market-summary benchmark, AI strategy reviewer scorecard, AI risk reviewer scorecard, hallucination audit, source-grounding report, model comparison matrix, RAG-over-research benchmark, local-vs-frontier market research report.
Research Question: Can risk dashboards and control systems expose the right information, alerts, limits, and intervention points to reduce market-system failure and prevent uncontrolled behavior?
Why It Matters: A market system without risk controls is not serious. Risk controls must be visible, enforceable, and reviewable, monitoring drawdown, exposure, daily/weekly loss limits, broker connectivity, and kill switches.
Systems / Tools Involved: Risk dashboards, trading dashboards, Grafana-style dashboards, custom web dashboards, P&L panels, exposure panels, drawdown monitors, order-state panels, data feed monitors, broker status checkers, alert modules, kill switch logic, circuit breaker blocks, pre-trade rules, post-trade reconcilers.
Benchmark Method: Risk-state visibility scoring, alert usefulness scoring, false alert rate, missed alert rate, time-to-intervention tracking, drawdown breach detection, exposure breach checks, kill switch trigger tests, circuit breaker simulation, failed-order simulations, runaway automation loops.
Expected Outputs: Risk dashboard design report, risk control checklist, kill switch protocol, circuit breaker design memo, exposure control map, drawdown alert plan, operator intervention protocol, risk dashboard mockup, risk telemetry plan, incident review template, paper-to-live risk-readiness checklist.
Research Question: What simulation and paper-trading environment is required before any market system could be considered for live-risk review?
Why It Matters: Backtests are not enough. Systems may backtest well and still fail when exposed to delayed data, broker disconnects, order rejections, partial fills, slippage, spread expansion, and latency. Simulation helps expose these problems before capital is at risk.
Systems / Tools Involved: Paper trading environments, broker paper accounts, exchange sandboxes, simulation engines, event-driven simulators, order lifecycle simulators, fill/cancel/rejection simulators, quote replay tools, trade print logs, latency/slippage simulators, fee models, strategy health dashboards, telemetry logs.
Benchmark Method: Paper-trading consistency review, simulated order lifecycle testing, broker disconnect simulation, data feed gap simulation, order rejection simulation, slippage and latency sensitivity testing, position reconciliation review, paper vs expected behavior comparison, telemetry completeness audit.
Expected Outputs: Simulation environment report, paper-trading monitor design, order lifecycle simulator spec, paper-to-live readiness checklist, telemetry completeness report, simulation failure-mode taxonomy, broker sandbox evaluation, paper-trading dashboard mockup, incident review protocol, escalation criteria memo.
Research Question: Can Source create clear, auditable maps of broker APIs, exchange APIs, FIX concepts, WebSocket streams, order lifecycle events, fills, cancellations, rejections, and reconciliation logic?
Why It Matters: Market systems fail at the edges. A model may work, but operational failure can occur when connecting to broker APIs, WebSocket streams, order gateways, FIX sessions, account permissions, rate limits, and API keys.
Systems / Tools Involved: Broker APIs, exchange APIs, FIX protocol concepts, WebSocket data streams, REST APIs, streaming quotes, streaming order state, account/order/position endpoints, trade history APIs, API key management, rate limit handling, broker adapters, execution gateways, reconciliation logic.
Benchmark Method: API documentation review, endpoint map creation, order lifecycle mapping, state transition testing, fill/cancel/reject scenario testing, rate limit checks, permission boundary audits, credential handling reviews, paper API testing, WebSocket reliability tests, disconnect/reconnect latency.
Expected Outputs: Broker connectivity map, exchange connectivity map, FIX concept map, WebSocket stream map, order lifecycle diagram, order state machine diagram, reconciliation checklist, broker API evaluation report, execution failure-mode taxonomy, secure credential handling checklist.
Research Question: Can Source create performance analytics and regime-analysis workflows that reveal strategy behavior honestly across market conditions, risk states, slippage, degradation, and drawdown periods?
Why It Matters: A single return number is meaningless. A serious strategy review requires Sharpe, Sortino, Calmar, max drawdown, drawdown duration, profit factor, expectancy, win rate, adverse/favorable excursion, slippage, turnover, exposure, and regime performance metrics.
Systems / Tools Involved: Python, pandas, NumPy, Polars, Jupyter, performance dashboards, risk dashboards, trade logs, backtest reports, paper trading reports, equity curves, drawdown charts, rolling metric dashboards, regime classification, volatility regimes, correlation analysis.
Benchmark Method: Metric completeness review, rolling performance analysis, drawdown duration analysis, regime split analysis, performance degradation detection, slippage and fee sensitivity review, trade distribution review, adverse/favorable excursion audits, volatility regime comparison.
Expected Outputs: Performance analytics report, regime analysis report, rolling performance dashboard, drawdown analysis memo, strategy degradation report, metric completeness checklist, slippage impact report, trade distribution report, performance dashboard concept.
Research Question: Can structured journaling, rule adherence, review loops, emotional-control protocols, and operator discipline systems reduce impulsive decisions and improve market-system operation?
Why It Matters: Even a good system can be ruined by poor operation. Market-system operators can fail through revenge trading, overtrading, ignoring drawdown limits, overriding rules impulsively, strategy hopping, and ignoring risk dashboards.
Systems / Tools Involved: Trading journals, decision logs, rule checklists, pre-trade checklists, post-trade review forms, drawdown review forms, daily/weekly review protocols, emotional-state logs, override logs, discipline scorecards, operator dashboards, AI journal/risk reviewers.
Benchmark Method: Rule adherence scoring, journal completeness review, override frequency analysis, revenge trading markers, overtrading markers, drawdown response review, pre-trade checklist compliance, emotional-state correlation, operator decision audits, AI journal review checks.
Expected Outputs: Trading discipline protocol, operator journal template, rule adherence scorecard, override review report, process adherence dashboard, trading psychology research memo, operator failure-mode taxonomy, AI journal reviewer workflow, daily/weekly review templates.
Research Question: Can logs, metrics, traces, alerts, dashboards, broker status, data-feed monitoring, latency monitoring, strategy health checks, and incident reviews make market systems more auditable and reliable?
Why It Matters: If a system cannot be observed, it cannot be trusted. Market systems need telemetry for data feed health, broker connectivity, API errors, WebSocket disconnects, order state, execution latency, strategy health, failed jobs, server health, and model calls.
Systems / Tools Involved: Grafana-style dashboards, Loki-style logs, Prometheus, OpenTelemetry, Tempo-style traces, InfluxDB, Mimir, Telegraf, model-call logs, token-cost logs, latency logs, broker connection logs, data feed health logs, order-state logs, execution traces, risk-limit alerts.
Benchmark Method: Telemetry coverage review, missing log detection, alert usefulness scoring, false alert rate, missed alert rate, time-to-detection, incident replay, latency analysis, broker disconnect detection, data feed freshness testing, failed job detection, risk alert simulation.
Expected Outputs: Trading telemetry architecture memo, observability coverage report, broker monitoring dashboard, data feed health dashboard, latency monitoring report, strategy health dashboard, risk alert plan, incident review template, telemetry gap analysis, alert taxonomy.
Infrastructure Layer

Premium Data & Market Systems
And Proprietary Research Environment.

Selected researchers may receive controlled, qualification-based access depending on project fit, approval, licensing, availability, cost, confidentiality, and research need.

Includes proprietary Source market research configurations, confidential workflows, internal research templates, strategy cards, risk review systems, custom benchmark environments, private pipelines, and experiment ledgers.

Source may provide selected researchers with controlled access to proprietary market research configurations, confidential workflows, internally developed harness systems, private research templates, custom model-routing setups, experimental dashboards, benchmark protocols, and Source-built evaluation environments. Some systems may be disclosed only after qualification, project assignment, and confidentiality review.

Includes Bloomberg Terminal, Bloomberg B-PIPE, Bloomberg Market Data Feed, LSEG Workspace, Refinitiv Workspace, FactSet Workstation, S&P Capital IQ, ICE Data Services, MSCI, Barra, Axioma, PitchBook, Preqin, and AlphaSense.

Selected researchers may work with premium or institutional-grade market-data environments, including terminal workflows, real-time feeds, historical tick data, exchange feeds, broker/exchange APIs, alternative data, and simulation-ready datasets where available, licensed, approved, and project-appropriate. Access to premium data, terminals, paid datasets, or institutional systems is not guaranteed and may depend on project scope, licensing, availability, qualification, and approval.
  • Direct exchange streams: Nasdaq TotalView, Nasdaq ITCH, Nasdaq Data Link, NYSE Integrated Feed, NYSE OpenBook, NYSE ArcaBook.
  • Derivatives & Commodities: Cboe DataShop, CME Market Data Platform, CME DataMine, OPRA options chain data, Eurex, ICE exchange feeds.
  • Reference feeds: Corporate action notifications, ETF constituent tracking indices, and reference mapping codes.
  • Quant APIs: Databento, Polygon.io, Massive market data, dxFeed, Intrinio, Tiingo, IEX Cloud, Financial Modeling Prep.
  • Historical tick pools: AlgoSeek-style historical databases, Twelve Data, Alpha Vantage, Finnhub, EOD Historical Data.
  • Streaming feeds: Exchange WebSockets, quote stream gateways, options chain APIs, and broker market data endpoints.
  • Regulatory & Filings: SEC EDGAR filings, analyst estimates, earnings call transcripts, economic calendars.
  • Macro economics: FRED databases, central bank publications, inflation datasets, interest-rate records.
  • Sentiment & Proxies: RavenPack-style news analytics, social sentiment tracking, Polymarket prediction feeds, weather datasets.
  • Aggregators: Kaiko, Coin Metrics, Glassnode, CryptoCompare, Amberdata, CoinGecko, CoinMarketCap.
  • Exchange integrations: Binance API, Coinbase API, Kraken API, OKX API, Bybit API order-book streams.
  • On-Chain trace: Wallet/entity labels, DEX liquidity statistics, stablecoin flow tracking, DeFi protocols, explorer logs.
  • Ingestion engines: Tick data processing, symbol normalization, timestamp alignment,timezone conversion routines.
  • Data engines: ClickHouse, DuckDB, TimescaleDB, InfluxDB, PostgreSQL, Parquet file archives.
  • Testing frameworks: Vectorized backtesting, event-driven backtesting, NautilusTrader engines, broker sandbox environments.

Includes brokerage APIs, exchange endpoints, FIX protocol sessions, WebSocket data streams, REST APIs, order routing concepts, execution adapters, and order state machines (cancellations, fills, rejections).

*Safety Boundary: Access to live credentials or live execution systems is not guaranteed.

  • Frontier configurations: ChatGPT/OpenAI, Claude, Claude Code, Gemini, DeepSeek, Qwen, Kimi, MiniMax.
  • Local runtimes: Ollama, vLLM servers, llama.cpp, quantized GGUF parameters, local embedding setups.
  • Agent architectures: Market research agents, strategy reviewer models, risk assessment assistants, RAG databases.

Fine-tuned LLMs, LoRA adapters, domain specialized financial assistants, SEC filings summarizers, earnings transcript models, and strategy critique setups.

Researchers may help evaluate whether custom-tuned, retrieval-grounded, or domain-specialized market research models outperform generic frontier models on Source-specific market systems tasks.

Includes research on hallucination-associated neuron detection, activation steering, Sparse Autoencoders interpretability, neuron suppression, model calibration checks, and grounding tests.

Source is interested in neuron-level hallucination mitigation research, including hallucination-associated neuron detection, activation steering, local model intervention experiments, groundedness evaluation, financial factuality testing, and reliability benchmarking. We make no claims that hallucinations are solved, that we possess non-hallucinating models, or that perfect factuality is guaranteed.

Agent orchestration structures, Claude Code workspaces, task run ledgers, vector registries, execution traces, model comparisons, and BuildGraph metrics.

Researchers may work inside custom Pi-style harness environments where AI agents are assigned, observed, logged, benchmarked, corrected, and compared across controlled market research and quant workflow experiments.
  • Vector databases: pgvector, Qdrant, Pinecone, Weaviate, Chroma index setups.
  • RAG features: Chunking parameter tests, hybrid search filters, semantic similarity, retrieval evaluation.
  • Grounding checks: Factuality scoring, citation correctness scripts, vector recall diagnostics.
  • Monitors: Grafana dashboards, InfluxDB time series, model token costs, order-state counters, latency logs.
  • Logs & Traces: Loki-style logs, Prometheus metrics, OpenTelemetry tracing, Tempo trace maps.
  • Incidents: System event registers, kill switch action logs, circuit breaker status displays.
  • Servers: Private Linux VMs, Windows remote RDP workspaces, dedicated high-core CPU servers, high-RAM boxes.
  • Tooling: Remote JupyterLab servers, SSH boxes, Docker/Podman containers, dataset storage nodes.
  • Distributors: Ray or Dask parallel configurations, Spark data processors, queue task handlers.

Composed settings linking premium data feeds, backtesting libraries, telemetry logs, risk panels, model routers, and research assistants.

The most valuable market research environment is not any single terminal, model, data feed, or backtesting tool. It is the composition of proprietary Source configurations, premium data, local inference, AI research agents, custom harnesses, RAG corpora, simulation systems, telemetry, dashboards, and human review into controlled experimental systems.
Proof-of-Work

Research Output
Comes First.

The Trading Systems Researcher lane is output-driven. We measure value by the clarity, correctness, and reproducibility of the scorecards, comparison reports, and maps submitted.

01

Data Feed Comparisons

Reports comparing market data vendors, terminal workflows, exchange feeds, APIs, alternative data, and on-chain datasets for research utility.

02

Pipeline Audits

Reports reviewing ingestion pipelines, database schemas, timezone conversions, normalizations, and query performance metrics.

03

Tick / OHLCV Audits

Reports testing whether tick files, OHLCV bars, order book depth files, and reconstructed quotes are aligned and simulation-ready.

04

Backtest Integrity

Reports identifying lookahead bias, survivorship bias, data leakage, parameter overfitting, and slippage/fee modeling errors.

05

Strategy QA Memos

Reports reviewing hypotheses, assumptions, signal logic, parameter sweeps, metrics coverage, and out-of-sample reproducibility.

06

AI Quant Evaluations

Reports checking whether model-generated market summaries, hypothesis lists, or strategy critiques contain hallucinations.

07

Risk Panel Reports

Reports checking drawdown visibility, exposure limits, alerting structures, and manual/automatic kill-switch readiness.

08

Simulation Memos

Reports evaluating paper-trading behaviors, broker sandboxes, latency slips, rejections, and escalation checklist readiness.

09

Broker API Maps

Reports diagramming broker API endpoints, exchange schemas, FIX tags, WebSocket streams, and position reconciliation checklists.

10

Performance Analytics

Reports calculating Sharpe, rolling drawdown duration, slippage ratios, correlation shifts, and model regime performance.

11

Trading Telemetry

Reports checking trace logs, metrics, alerts, queue health, connection states, failed jobs, and incident logs.

12

Proof-of-Work

Public or private artifacts demonstrating real research contribution, where mutually approved and appropriate.

Every formal research memo submitted to the lab is structured to include:

1. Research Question
11. Failure Cases
2. Target Hypothesis
12. System Limitations
3. Systems Tested
13. Risk Considerations
4. Tools Used
14. Screenshots or Logs
5. Data or Corpus Used
15. Reproducibility Notes
6. Market / Instrument Scope
16. Confidence Level
7. Method
17. Recommendations
8. Benchmark Criteria
18. Next Experiments
9. Assumptions
19. Reviewer Notes
10. Results
20. Access / Licensing Notes
Access Model

Access Is
Qualification-Gated.

We provide controlled, project-based access to research tools. The pipeline moves from public reviews to matching programs and deeper review.

01

Review

Candidate reviews parent lab layouts, quant programs, questions, and risk boundaries.

02

Apply

Candidate submits background, proof-of-work, availability, and research interests.

03

Vet

Source evaluates technical capabilities, documentation clarity, and risk awareness.

04

Match

Candidate matched to specific, scoped AI market systems research programs.

05

Access

Researcher receives scoped sandbox access matching the approved project scope.

06

Output

Researcher submits structured validation reports, scorecards, comparisons, and maps.

07

Review

Strong candidates evaluated for paid collaboration, fellowships, or future research tasks.

Selected researchers may receive controlled, qualification-based access to Source market research environments, proprietary configurations, custom harnesses, private research templates, model workflows, benchmark protocols, dashboards, data workflows, and infrastructure depending on project fit, approval, licensing, confidentiality, and availability.

Not all infrastructure is publicly described. Certain technologies, configurations, internal benchmarks, private datasets, premium data relationships, and experimental systems are intentionally held back from the public page and may only be discussed with qualified candidates after review.

Access to premium data, terminals, compute, proprietary systems, broker APIs, exchange feeds, local inference environments, frontier models, or paid systems is not guaranteed and may depend on project scope, availability, cost, licensing, qualification, and approval.

Boundaries

What This Research Lane
Is Not.

We operate a serious applied systems research space. Verify these non-guarantees prior to submitting consideration logs.

No Signals Service

This is not a signal group, trading course, prop-firm passing tool, or investment promotion. There are no pips, profit margins, or buy/sell setups offered.

No Open Access

We do not provide free API tokens, Bloomberg credentials, or FactSet access for personal trading. Resources are locked to scoped research deliverables.

No Profit Guarantees

Testing market pipelines does not guarantee live trading permissions, account funding, or payouts. Vetting is strictly educational and systems-oriented.

Research Output Comes First

This page is not offering unrestricted access to models, terminals, data feeds, servers, broker APIs, dashboards, or proprietary Source systems. Access is tied strictly to qualification, trust, project fit, confidentiality, licensing, availability, resource cost, clear scope, risk boundaries, and useful research output.

Application Preview

What Applicants Should Be
Ready To Answer.

We vet candidates systematically. Review these targeted questions to prepare your research log submissions.

1. General & Markets

  • • Why apply to Source Research Lab as a Trading Systems Researcher?
  • • What interests you most: data, agents, backtests, risk, or simulation?
  • • What markets have you studied or worked with?
  • • What proof-of-work can you show?
  • • How many hours can you commit?

2. Tech & Tools

  • • Have you used Python, pandas, NumPy, Polars, or Jupyter?
  • • Have you worked with tick, OHLCV, options, or crypto data?
  • • Have you used Bloomberg Terminal, LSEG Workspace, FactSet, or Databento?
  • • Have you used vectorbt, backtrader, or QuantConnect/LEAN?
  • • Have you worked with broker APIs, FIX, WebSockets, or paper environments?

3. Research Judgment

  • • What does backtesting integrity mean, and what fakes them?
  • • What is a critical data-feed or API connectivity failure mode?
  • • How would you test if a market data feed is reliable?
  • • How would you test if a model-generated market summary hallucinated?
  • • How would you design a paper-to-live readiness checklist?

4. Boundaries & Vetting

  • • Are you comfortable working with confidential, staged systems?
  • • Are you willing to write structured reports rather than notes?
  • • Are you comfortable receiving direct feedback?
  • • Do you understand that access, paid work, compute, and returns are not guaranteed?
Support Desk

Frequently Asked
Questions.

Review clarification details concerning testing scopes, requirements, and risk boundaries.

No. This is a selective research and benchmark lane inside Source Research Lab. Strong candidates may be considered for deeper involvement, paid collaboration, or future Source tasks, but nothing is guaranteed.
No. The Source University Trading Systems Program is designed to train candidates. This page is for the research lane that evaluates the systems, data workflows, backtests, simulations, dashboards, risk controls, and infrastructure behind that training.
No. This is not investment advice. It is research-focused, benchmark-focused, simulation-first, paper-trading-first, and risk-aware.
No. This is not a signals service. The page does not promise trade alerts, entries, exits, profits, account growth, funding, or live-trading access.
No. Profits are not guaranteed. Trading involves substantial risk. AI, premium data, backtesting, machine learning, simulations, and dashboards do not guarantee profitable decisions.
No. Funding is not guaranteed. This is not a prop-firm passing program, funding program, or account-growth offer.
No live trading is guaranteed. The research posture is simulation-first and paper-trading-first. Any live-risk consideration, if ever relevant, would require separate approval, risk review, and controls.
No. A PhD is not required. Strong technical ability, data awareness, research discipline, structured documentation, proof-of-work, risk awareness, and seriousness matter more than titles.
Market experience is useful, especially for data feeds, backtesting, risk dashboards, broker APIs, and simulations. However, strong data, engineering, ML, AI, or systems research ability may also be relevant.
Coding ability is strongly useful, especially for market data pipelines, backtesting, simulation, broker API research, dashboards, notebooks, telemetry, local inference, and AI research workflows.
Depending on fit, you may help evaluate market data feeds, backtesting workflows, simulation environments, paper-trading monitors, broker/exchange connectivity maps, risk dashboards, AI market research agents, local inference labs, premium data workflows, telemetry systems, or proprietary Source market configurations.
Possibly, depending on project scope, licensing, availability, qualification, and approval. Access to Bloomberg, LSEG, FactSet, ICE, exchange feeds, paid APIs, or premium datasets is not guaranteed.
Possibly. The lane includes broker/exchange connectivity mapping, API documentation review, WebSocket stream mapping, paper API testing, order lifecycle diagrams, and reconciliation logic. Access to live credentials or live systems is not guaranteed.
Possibly. The lane is simulation-first and paper-trading-first. Researchers may help evaluate paper workflows, simulated order lifecycle behavior, broker errors, data-feed gaps, risk-limit breaches, and readiness checklists.
Possibly. Selected researchers may evaluate market research agents, strategy reviewer agents, risk reviewer agents, trade journal reviewer agents, AI-generated research briefs, RAG-over-market-research workflows, and model-routing systems.
Possibly. Selected researchers may work with local models, private inference endpoints, custom-tuned market research models, retrieval-grounded market agents, or local-vs-frontier benchmark tasks where available and approved.
Possibly. Public proof-of-work, case studies, testimonials, or public research summaries may happen only where mutually approved and appropriate. Public exposure is not guaranteed.
Possibly. Strong candidates may be considered for paid collaboration, deeper research involvement, or future Source tasks if qualified. This is not guaranteed.
No. Access is controlled, project-scoped, and subject to availability, cost, licensing, approval, qualification, and confidentiality boundaries.
Expected outputs may include data-feed comparison reports, market data pipeline audits, tick/OHLCV validation reports, backtesting integrity reports, strategy QA reports, AI quant research evaluations, risk dashboard reviews, simulation reports, broker connectivity maps, performance analytics reports, telemetry architecture memos, and proof-of-work artifacts.
Applicants should apply for Trading Systems Research review and provide background, proof-of-work, research interests, technical experience, market-systems experience, availability, and willingness to produce structured outputs.
Trading Systems Research Review

Apply For
Trading Systems Research Review.

If you are serious about helping test, benchmark, document, and improve advanced market data pipelines, premium data workflows, backtesting systems, simulations, broker/exchange maps, risk dashboards, AI market research agents, telemetry systems, local inference labs, and proprietary Source market research environments, you may apply for Trading Systems Research review.

Trading Systems Research Qualification-Gated Outputs Matter
△ Apply for Trading Systems Research Review Back to Source Research Lab → View AI Systems Research Lane
Access, paid collaboration, public exposure, premium data, compute, proprietary systems, model workflows, local inference labs, custom harnesses, broker/exchange workflows, live-risk systems, and future Source involvement are not guaranteed. They are qualification-based, project-scoped, and subject to approval, availability, licensing, confidentiality, risk boundaries, and research fit.