AI ENGINEER •DATA SCIENTIST •SOFTWARE ENGINEER

Jack Lutz

I Build Systems that Scale, and Answer Questions that Matter.

Las Vegas, NV • Stats & Data Science | Mathematics @ UCLA '27

PythonRSQLTypeScriptJavaScriptGoCC++SwiftHTMLCSSSvelteLaTeXPythonRSQLTypeScriptJavaScriptGoCC++SwiftHTMLCSSSvelteLaTeXPythonRSQLTypeScriptJavaScriptGoCC++SwiftHTMLCSSSvelteLaTeXPythonRSQLTypeScriptJavaScriptGoCC++SwiftHTMLCSSSvelteLaTeX

GitpandasNumPymatplotlibTidyverseAWSDockerLinuxFastAPIPostgreSQLSQLiteRedisSupabasePyTorchScikit-LearnStatsmodelsFAISSLangChainOpenRouterHugging FaceGoogle OR-ToolsCUDAMCPReactChrome Extension MV3StreamlitCursorClaude CodePrompt EngineeringRetrieval Augmented GenerationVector DBGitpandasNumPymatplotlibTidyverseAWSDockerLinuxFastAPIPostgreSQLSQLiteRedisSupabasePyTorchScikit-LearnStatsmodelsFAISSLangChainOpenRouterHugging FaceGoogle OR-ToolsCUDAMCPReactChrome Extension MV3StreamlitCursorClaude CodePrompt EngineeringRetrieval Augmented GenerationVector DBGitpandasNumPymatplotlibTidyverseAWSDockerLinuxFastAPIPostgreSQLSQLiteRedisSupabasePyTorchScikit-LearnStatsmodelsFAISSLangChainOpenRouterHugging FaceGoogle OR-ToolsCUDAMCPReactChrome Extension MV3StreamlitCursorClaude CodePrompt EngineeringRetrieval Augmented GenerationVector DB

Mathematical StatisticsProbability TheoryStatistical ComputingStatistical OptimizationExperimental DesignData Analysis & RegressionAbstract Linear AlgebraDifferential EquationsVector CalculusGame TheoryData Structures & Algorithms IData Structures & Algorithms IIMathematical StatisticsProbability TheoryStatistical ComputingStatistical OptimizationExperimental DesignData Analysis & RegressionAbstract Linear AlgebraDifferential EquationsVector CalculusGame TheoryData Structures & Algorithms IData Structures & Algorithms IIMathematical StatisticsProbability TheoryStatistical ComputingStatistical OptimizationExperimental DesignData Analysis & RegressionAbstract Linear AlgebraDifferential EquationsVector CalculusGame TheoryData Structures & Algorithms IData Structures & Algorithms IIMathematical StatisticsProbability TheoryStatistical ComputingStatistical OptimizationExperimental DesignData Analysis & RegressionAbstract Linear AlgebraDifferential EquationsVector CalculusGame TheoryData Structures & Algorithms IData Structures & Algorithms II

PythonRSQLTypeScriptJavaScriptGoCC++SwiftHTMLCSSSvelteLaTeXGitpandas PythonRSQLTypeScriptJavaScriptGoCC++SwiftHTMLCSSSvelteLaTeXGitpandas

Skills

Tech Stack

Background

Experience

UCLA Department of Mathematics

Researcher

Spring 2026

Los Angeles, CA

• Researching statistical arbitrage in cryptocurrency markets: constructing normalized spread series across liquid trading pairs and testing for mean reversion via cointegration analysis with walk-forward evaluation.
• Built a survivorship-bias-free dynamic universe spanning 100+ liquid spot pairs, rebalancing daily on volume and spread thresholds.
• Tested mean-reversion hypotheses via Engle-Granger and Johansen cointegration on rolling windows; quantified half-life of mean reversion to size positions.

Lutz Consulting Group

Founder

Nov 2024 - Present

Las Vegas, NV

• Analyzed 138 months of client transaction data; built Gradient Boosting forecasting model with engineered lag, rolling, and seasonal features, achieving 11.5% MAPE, a 58% improvement over naive baseline on 24-month holdout, uncovering $550K+ in dead stock for liquidation.
• Engineered computer vision pipeline using GPT-4 Vision and AWS S3 to extract and classify 40,000+ product tags; built Streamlit review dashboard backed by PostgreSQL, enabling staff to reconcile full inventory in one weekend.
• Shipped end-to-end ML pipelines (data ingest → model → review UI → action) for non-technical clients; owned scoping, modeling, validation, deployment, and training.

New York Life

Data Analyst Intern

June 2025 - Aug 2025

New York, NY

• Built RAG-enabled underwriting chatbot prototype using company underwriting manual as knowledge base, demonstrating feasibility of LLM-powered policy guidance.
• Designed controlled experiments measuring LLM sycophancy: varied prompt authority framing across multiple models, measured response drift toward user-implied conclusions, and produced the team's first empirical model behavior finding that drove production prompt engineering policy.
• Finalist in company-wide AI Innovation Competition for prototyping RAG-enabled chatbot plugin for Tableau dashboards, querying 10+ years of nationwide data with office-level granularity.

Work

Projects

UCLA Dining Recommendation System

Full data pipeline and recommender system to help college students hit their macro goals on their meal plan.

Python FastAPI PostgreSQL Google OR-Tools Go

• Go scraper ingesting 38,000+ menu items daily into PostgreSQL
• Google OR-Tools CP-SAT solver generates macro-optimized meal plans in 52ms average solve time
• FeedMe app built on top of this pipeline currently in beta testing for UCLA students

Code not public

Quantitative Research Backtester & Platform

Full-stack quantitative research platform: strategy IDE, real-time charting, and MCP AI integration layer on top of a co-developed C++ backtesting engine with custom DSL and per-fold walk-forward validation.

Python C++ TypeScript Linux Redis AWS MCP

• Seven-figure AUM allocation from institutional partner (Mazer Group); live pilot generating positive alpha vs S&P 500
• Data pipeline spanning equity, crypto, macro, and fundamental data across 13,000+ instruments
• 27 MCP-orchestrated tools with model-tiered routing; L0/L1/L2 context hierarchy + prompt caching cut per-run cost from ~$10 to ~$0.50

Code not public

Interactive Demo

DSA from Scratch

Implemented data structures and algos from scratch for practice, and turned this into a visualizer for sorting algos, and a maze generator and solver. Check out the interactive demo.

Python SvelteKit TypeScript

• 7 sorting algorithms
• Maze generation & 4 solving algorithms
• Real-time step-by-step visualization

GitHub Repo Try Demo

Claude-o-Meter

Native menu-bar (macOS) and system-tray (Windows) app showing live Claude Pro/Max 5-hour and weekly subscription usage. Pulls straight from the same Anthropic endpoint the desktop app's Settings → Usage page uses, so percentages match exactly.

Swift Rust

• Cross-platform: Swift menu-bar app on macOS, Tauri 2 + Rust tray app on Windows; both shipped via signed installers and CI release workflow

GitHub Repo

tinynccl: RDMA PyTorch Backend

From-scratch C++ GPU collective-communication library on libibverbs/softRoCE, exposed as a custom PyTorch c10d::Backend. Used to train a 10.7M-parameter char-level GPT on TinyShakespeare via DDP through code I wrote.

C++ CUDA libibverbs PyTorch

• All-reduce kernel validated to 4MB buffers under 1000-iteration stress runs
• Custom c10d::Backend lets DistributedDataParallel route gradient sync through the library transparently
• Trained 10.7M-param GPT across two ranks with bit-exact cross-rank weight consistency

GitHub Repo

Interactive Demo

ML from Scratch

Implemented core ML algos from scratch in Python with visuals and explanations to test my understanding. Used Elements of Statistical Learning for foundations.

Python NumPy Streamlit R

• 16 models built from scratch using only NumPy and math
• Interactive demo on Streamlit
• Models rigorously benchmarked against scikit-learn and PyTorch

GitHub Repo Streamlit Page

Live Demo

Wikipedia Speedrun Benchmark

A benchmark comparing AI agents playing the Wikipedia Speedrun game - navigating between Wikipedia articles using only page links in as few clicks as possible.

Python Hugging Face Docker Embeddings

• 15 AI agents benchmarked across 35 problems (420+ games)
• Embedding-based agents achieve 94% win rate at zero cost
• Interactive demo lets you play or watch AI agents compete

GitHub Repo Hugging Face Space

Extension Available

Web Scraper Chrome Extension

A Chrome extension with AI-powered data extraction and intelligent cleaning, supporting multiple export formats.

TypeScript JavaScript Chrome Extension MV3

• AI-powered data extraction and cleaning
• CSV, XLSX, JSON export support

GitHub Repo Extension Download

Claude Code Framework

Structured workflows for Claude Code. I developed this system when using Claude Code in large codebases, and it is inspired by Boris Cherny's Claude Code workflow.

Claude Code CLI Plugin

• Persistent codebase context across sessions
• Enforced review checkpoints and decision tracking
• Built-in custom agent orchestration for reliable implementation, review, and testing

GitHub Repo

Contact

Get in touch

Send Email