01 / CRYPTO RESEARCHCrypto Alpha Research
Research workflow covering a 446-asset universe, market regime intelligence, causal validation, multi-stage screening and model governance.
PYTHON · QUANT RESEARCH · AI AUTOMATIONQUANTITATIVE RESEARCH · APPLIED AI · SYSTEMS ENGINEERING
Engineering
conviction from
market noise.Research first.
Alpha follows.
I'm Mochamad Edgar Erlangga, a Python engineer building quantitative research systems, crypto backtesting, and AI automation. I turn data into evidence and ideas into practical, testable software.
Independent research & engineeringAvailable worldwide · Remote
01
SYSTEMATIC
TRADING
RESEARCH
TRADING
RESEARCH
446CLOSED TRADE RECORDS FASTGRAB / ASOF V3
60.31%HISTORICAL WIN RATE 2024–2026
1.548PROFIT FACTOR REPORTED CLOSED PNL
96,849CANDLES ANALYZED SOLUSDC · 15-MINUTE
QUANTITATIVE RESEARCH✳PYTHON BACKTESTING✳CAUSAL VALIDATION✳AI AUTOMATION✳PRODUCTION ENGINEERING
01 / FEATURED RESEARCHA strategy is a hypothesis.
A strategy is a hypothesis.
Evidence is the product.
Transparent case-study analytics from a private crypto strategy research system. Proprietary entry logic and parameter files remain confidential.
FASTGRAB / ASOF V3SYSTEMATIC CRYPTO TRADING RESEARCH
HISTORICAL REPLAY · SOLUSDC · 15M
QUANTITATIVE PERFORMANCE / 2024–2026
Based on monthly closed-trade summariesHistorical PnL, under the microscope.
Cumulative reported closed-trade PnLUSD · monthly close groupings · not a reconciled equity curve
Net PnLHistorical simulation only. Compounding changes dollar position sizes over time.
Monthly net PnL · 2026USD millions
Profit factor by yearAggregate reported values
Trade direction splitClosed records
BUY 196SELL 250
Long win rate 61.73%Short win rate 59.20%
TIME-SLICED PERFORMANCE
Review the months, not just the endpoint.
2026 monthly net PnLUSD · reported realized trade PnL within the historical backtest
Positive NegativeTRADE-LEVEL SUMMARY
Aggregated ledger output, not fabricated fill historyDistribution and direction.
Directional comparison
SideTradesWin rateProfit factor
BUY / Long19661.73%1.592
SELL / Short25059.20%1.419
Dollar PnL between sides is affected by compounding and differing position sizes. This does not establish directional causal superiority.
Exit types
Limit maker take profit269
Market taker stop177
The engine reports 450 executed trades while 446 eligible closed records were summarized. The remaining difference must be reconciled before treating the record as execution-complete.
DURABILITY & COSTS
Annual report · 2026 YTD through OctoberWhere the evidence gets harder.
Annual profit-factor trajectory
2024 · win rate62.13%169 closed trades
2025 · win rate62.01%179 closed trades
2026 · win rate54.08%98 closed trades
Sum of recorded fees$2.41MReported trade record field
Win rate and profit factor declined from 2024 to 2026. No independently validated mark-to-market max drawdown is published.
FULL DISCLOSURE
What the research does—and does not—prove.
What was measured
FastGrab ASOF V3 was replayed on 96,849 SOLUSDC 15-minute OHLCV bars covering January 2024 to October 2026, using a frozen strategy configuration and state policy. The simulation starts from $10,000 and compounds after individual wins and losses.
Its engine reports a hypothetical ending balance of $38,313,177.82. This is not a live return and is not presented as a realistic deployable capital forecast.
Unresolved validation items
- 450 reported engine executions versus 446 audited closed records.
- Six recovered candles are not fully cross-source verified.
- Liquidity, funding, costs, capacity and slippage at scaled size require further testing.
- Independent OOS and reliable mark-to-market drawdown are not yet demonstrated.
No proprietary source code, private model weights, exact entry filters or strategy parameter files are distributed through this portfolio.
Historical simulation · public aggregated data · provisional
i Historical simulation — not live returns. Published statistics are research outputs with known validation limitations.
02 / SELECTED PROJECTS
Beyond the backtest.
Research and engineering systems built across finance, automation and web software.
01 / CRYPTO RESEARCH
02 / AI AUTOMATIONAI Automation Systems
LLM-powered intake workflows, RAG applications, transcription and automated content/video-processing systems.
03 / DEVELOPMENTWeb Products & Integrations
Full-stack platforms, business workflows, API integrations and production deployment from prototype to launch.
03 / WHAT I CAN BUILD FOR YOUResearch-grade thinking.
Research-grade thinking.
Production-ready delivery.
Focused technical services for traders, quantitative teams and digital-first businesses.
Strategy Backtesting
Turn explicit trading rules into reproducible backtest pipelines with fees, trade logs, assumptions and performance metrics.
Discuss a project ↗Quant Research Engineering
Market-data validation, walk-forward workflows, causal audits, fast research infrastructure and repeatable analysis.
Discuss a project ↗AI & Software Automation
Python applications, LLM integrations, dashboards, APIs and reliable Linux/Docker production deployments.
Discuss a project ↗04 / THE PERSON BEHIND THE SYSTEMSDeveloper by craft.
Developer by craft.
Researcher by habit.
I build quantitative research tools, practical AI automation and full-stack applications. My long-term goal is to establish a proprietary quantitative trading research firm grounded in sound experimentation and disciplined risk management.
EXPERIENCEIndependent Quant Research & Engineering
Python backtests · Crypto market data · Applied AI
FULL-STACKJagat Sejahtera Abadi
Production web applications and integrations
AI EVALUATIONOutlier AI
Explore my GitHub ↗Language model quality and evaluation work
05 / BUILD SOMETHING MEANINGFULHave a research problem
Have a research problem
worth solving?
Send a brief on the market, strategy, workflow or system you want built. I'll help define a focused scope and deliverables.
Start a conversation ↗