Order Router & Execution Engine
Low-latency order routing service with risk checks and broker abstraction across Alpaca, IBKR, and Binance.
I'm Hasan Javed, a senior full-stack & AI engineer with 9+ years shipping fintech platforms, LLM-powered apps, and high-throughput APIs in Python, Go, and Rust. 50+ projects delivered, including trading systems that have moved $80M in 2025 alone.
from fastapi import FastAPI, Depends from app.exec import Router, Risk from app.brokers import alpaca, ibkr app = FastAPI(title="order-router") @app.post("/v1/orders") async def place(o: Order, r=Depends(Risk)): await r.check(o) # 38ms p99 fill = await Router(o).execute() return { "id": fill.id, "avg_px": fill.px }
I'm a senior full-stack engineer focused on the finance sector. Over the last nine years I've built fintech platforms with Stripe & Plaid, AI-powered apps on OpenAI / Gemini / Claude / Grok, and high-performance APIs serving millions of requests.
I work end-to-end — product discovery, architecture, implementation, infra, and post-launch ops. Most of my work lives somewhere in the stack between Python / Go / Rust services, React / TypeScript front-ends, and AWS (Lambda, ECS/EKS, RDS, AppSync).
When the brief is fuzzy, I help shape it. When the deadline is tight, I cut scope honestly. I write code my clients can keep maintaining after I hand it over — that's the part I care most about.
Most engagements pull from two or three of these. I'm happy to scope something tight, audit an existing system, or own a long-running platform from architecture through ops.
REST & GraphQL services built to scale, with clean docs and observability from day one.
Pragmatic AI features that actually move metrics — not demos.
Money movement, accounting, and compliance done right the first time.
Backtested strategies, risk-aware execution, and broker integrations.
Models that survive contact with real markets — and Parquet at scale.
End-to-end product builds, from MVP to a system you can hand to a team.
AWS-first, Docker-native, observable by default.
Browser extensions, Figma plugins, and quiet little tools that save hours every day.
I'm fluent across the stack and pick tools to fit constraints — team, deadlines, ops, latency budget — not the other way around.
Started shipping client work in 2016. Today I focus on fintech, AI/LLM systems, and high-throughput services in Python, Go, and Rust.
Began full-time freelance work building Django and Laravel platforms for SMBs — my first taste of owning end-to-end product delivery.
Moved deep into fintech: payment integrations (Stripe, PayPal, Square), accounting APIs (QuickBooks, Xero), and reporting pipelines on Parquet/Arrow.
Designed and ran multi-tenant systems on AWS Lambda, ECS, EKS, and AppSync — observability, IaC, and zero-downtime deploys baked in.
Built backtesting and live execution stacks across Backtrader, Zipline, QuantConnect, and MT5/MQL5 — integrated with Alpaca, IBKR, and Binance.
Shipped AI features on OpenAI, Claude, Gemini, and Grok — agents, RAG, evals, and custom orchestrations on n8n and bespoke runtimes.
Architected and shipped an AI-powered content platform that now serves over ten thousand daily active users in production.
Delivered execution and risk infra that moved $80M of live volume in Q1–Q3, with a 60% reduction in client-side data processing time.
A slice of recent client and product work — mostly in fintech and AI. Details under NDA where applicable.
Low-latency order routing service with risk checks and broker abstraction across Alpaca, IBKR, and Binance.
Multi-tenant AI content platform with model fan-out, evals, and a custom orchestration runtime.
Parquet/Arrow-backed reporting layer with Plaid + Stripe ingestion and reconciliation pipelines.
AI SaaS that helps individuals make 10X progress in their careers — resume scoring, job match, interview prep.
Strategy backtesting & walk-forward analysis layer over Backtrader, Zipline, and QuantConnect.
Bidirectional sync engine for QuickBooks, Xero, Wave, and AccountEdge with conflict resolution.
Lightweight runtime for tool-using LLM agents with model fan-out, retries, and structured-output evals.
Webhook bridge translating TradingView signals into broker-native orders with bank-side risk checks.
Internal-tooling plugins that automate handoff between design and engineering — saves hours weekly.
Nine years of building quantitative & ML-driven trading systems — from signal research and backtesting to low-latency execution and 24/7 live ops. Not just notebooks: production pipelines that route real orders, with risk, audit and failover wired in from day one.
Alpha hunted from price, order-flow, alt-data and LLM-tagged news. Hypothesis → notebook → paper → live, with a kill criterion written before the first backtest.
Low-latency order routers in Rust & Go. Smart routing across venues, iceberg / TWAP / VWAP slicing, fee-aware fills. p99 pre-trade risk under 40ms.
Gradient-boosted trees, LSTMs and transformer-based regime classifiers for short-horizon forecasting. Feature stores, leakage guards, cross-asset signals.
Walk-forward validation, purged K-fold, tick-resolution replay. Slippage, fees and market impact modelled honestly — no curve-fits to Sharpe-theatre.
Pre-trade checks, position limits, drawdown kill-switches, audit trails. Built for real money, not paper — the plumbing you don't see until you need it.
Observability, alerting, broker failover, cold-standby replicas. The model is 20% of the work — keeping it running without drift or outage is the other 80%.
Outcomes from recent engagements. Architecture diagrams, post-mortems, and reference calls available under NDA.
Order routing & execution infra moved $80M in Q1–Q3 2025.
AI-powered content platform serving 10K+ DAU in production.
Fintech reporting dashboard cut client-side processing time by 60%.
High-throughput APIs sustained above one million requests per day.
Names redacted for confidentiality.
Reference calls available on request.
“Hasan delivered a trading system that beat our throughput targets and shipped on time. He owns the work end-to-end — rare at his rate, rarer still at the level of polish.”
“Code we could actually keep maintaining after handover. Tests, docs, a clean deploy story, and an ops runbook. Best contractor we've worked with in three years.”
“He didn't just build features — he asked the product questions our team kept ducking and saved us from two expensive wrong turns in the first month.”
Every project starts with a free 30-min scoping call. After that, pick the structure that lowers your risk — not mine.
All tiers include: NDA + IP assignment · clean docs · test coverage · post-launch handover
The questions I get most often before a kickoff. Anything missing? Drop it in the contact form — I'll answer and add it here.
Send a short brief with your goals and timeline. I respond within 24 hours with a clear next step — whether that's a call, a proposal, or pointing you to someone better suited.
Also open to senior / staff full-time roles — remote, EU/US time zones. Mention it in the brief.