Work Authorisation
Singapore Citizen, eligible for H-1B1 US Visa, HPI UK Visa and the TTPS HK Visa with no petition required.
Education
- Optimisation, Stochastic Analysis, Monte Carlo Methods, Game Theory, Operations Research, Machine Learning, Computational Math, Data Science in Finance, Distributed Computing.
Projects
- 1st Place, Google DeepMind × Cactus Compute Hackathon 2026 (Solo, 90 entrants): Optimised FunctionGemma on-device inference with custom query decomposition and confidence routing to Gemini Flash. 0.96 F1, 60% on-device.
- 2nd Place, Google Gemini 3 Hackathon 2026 (Solo, 189 entrants): Contour — Full-stack app (FastAPI, Three.js, Gemini Live API) that converts 2D topo maps into flyable 3D terrain with voice tour guide. (5K+ likes)
- Shipped:
- Teletable: Published Mac menubar app serving live sports data (300+ users, Swift).
- Notchi: MacBook notch tamagotchi for Claude Code sessions (WIP, Swift, Open Source).
- OpenPPT: Model-agnostic LLM-powered slide generator (WIP, TypeScript, Open Source).
- KeepIt (VisionOS Game, Swift), AlgoTrade (ML Stock Selection), BeamSolve (FEM Calculator), Housing Value Prediction (KRR/MPI), Air Pressure Prediction (Apache Spark), Polymarket Whales (Telegram, Polymarket API). More at skruban.com
Skills
- Languages: Python, PostgreSQL, Swift, JavaScript, TypeScript, VBA, HTML/CSS, Haskell
- Technologies: NumPy, Pandas, PyTorch, REST APIs, FastAPI, Git, Bash, Docker, CI/CD (GitHub Actions), AWS
- AI/Agents: Claude Code, Codex, MCP, LangGraph
- Finance: CFA Level I (2022), Bloomberg Terminal
Experience
- Embedded with 10 CAD/AUD/FX traders as primary technical liaison during live market hours, triaging issues across data feeds, Excel tooling, and performance bottlenecks to minimise trading disruptions.
- Scoped requirements with AUD desk traders, built a VBA blotter system integrating Bloomberg API and proprietary data sources to track PnL and risk metrics, cutting calculation time by 50% while consolidating 20+ tabs to 12.
- Researched cubic spline interpolation methods for interest rates curve construction, delivering technical presentation to APAC senior traders on tradeoffs between risk locality, pricing accuracy, and curve smoothness.
- Received full-time offer to join APAC Interest Rates Quant team.
- Owned the coal, iron ore & metals data platform end-to-end for 50+ global clients: gathered requirements, prioritised backlog, wrote specs, and coordinated Engineering and Data Ops through full delivery lifecycle.
- Diagnosed and resolved production data issues live on client calls, feeding insights back into backlog prioritisation and product improvements — the primary technical point of contact for client escalations.
- Led platform expansion into new trading regions: engineered and deployed web scrapers and ETL pipelines (Python, REST APIs) capturing OCR bills of lading, port/customs API feeds, and broker data.
- Partnered with Data Science team to scope and deliver Random Forest models for commodity prediction; curated training datasets that improved grade and volume prediction accuracy by over 20%.
- Built automated Python validation system comparing Kpler data against third-party sources, eliminating 2 man-days of manual QA monthly.
- Led assessment of AI, Blockchain, and RPA technologies to develop a 5-year tax technology roadmap; managed external consultants to deliver automated travel tracking system for global tax compliance.
- Managed the entry-into-service of RR Trent engines across 7 APAC airlines, coordinating multi-country stakeholders to streamline procurement pipelines for on-time delivery.