jlitven(1)

Experience

WORK

LIQUIDITY LOCK · Florida, USA 2020 – present

Co-Founder, Head Quant & Head of Development

Co-founded with former LMRKTS colleagues to bring portfolio-style, sessions-based matching to market risk transfer. Participants describe what they want the way they'd brief a colleague; the engine finds the best trade set for everyone at once, confidentially and with a clear audit trail.

  • Designed and built the matching and optimization engine using a modern algebraic modeling language, matching complex orders at fill rates orders of magnitude higher than mid-market matching while adhering to participants' risk limits.
  • Built the backend of the trading platform: complex order types, curves and parameters, with sessions completing in minutes.
  • Lead a 3-person backend team; hired and mentored two junior developers.
  • Built a benchmark framework to measure and audit pipeline performance, then optimized core functionality by 50%.
  • Automated CI/CD and deployments to testing, staging and production with GitHub Actions and NSSM; productionized database migrations with Flyway.
  • Created nightly end-to-end cron tests of the full pipeline, from parsing data to database reports, logging outputs to MongoDB.
  • Built Claude skills that benchmark new pull requests, automate code review and integrate with Shortcut via MCP.
LMRKTS · New York, USA 2017 – 2020

Senior Quantitative Developer

  • Worked on the core FX compression algorithm, which eliminated over $4 trillion in gross FX notional exposure. Try a toy version →
  • Designed and implemented Bonsai, a live multilateral optimization service (Gurobi) that reduced the CCAR largest-counterparty exposure of 5 live clients.
  • Led firm-wide engineering-quality work: Git, Jira-based agile development, and unit, integration and regression testing across all projects.

EDUCATION

UNIVERSITY OF BRITISH COLUMBIA · Vancouver2010 – 2013

Master of Science, Computer Science

  • Developed a GPU optimization algorithm (CUDA) for simulating contact between highly deformable objects. By exploiting sparsity and problem structure, I cut the problem from thousands of variables to a handful and made simulations 3–4× faster. In practice: lots of virtual bunnies colliding.
UNIVERSITY OF WATERLOO · Waterloo2005 – 2010

Bachelor of Mathematics

  • Honors with Distinction, GPA 87%.
NYC DATA SCIENCE ACADEMY · New York2016

Data Science Certificate

  • Built a Spark collaborative-filtering recommender trained on 50 million ratings. Blog posts.

SKILLS

optimization
mathematical optimization, optimization modeling, Gurobi, CPLEX, algorithm design, benchmarking
python
gurobipy, pandas, NumPy, SciPy, FastAPI, SQLAlchemy, Streamlit, pytest
languages
Python, Java, SQL, R, Matlab
data & apis
PostgreSQL, MongoDB, GraphQL, REST
devops
CI/CD, GitHub Actions, NSSM, Flyway, Alembic, Git, end-to-end testing
ai tooling
Claude Code & agents, Cursor, MCP servers, LLM APIs
offline
language learning, dancing, strength training