Jason Bokinz
Full Stack Software Engineer
I graduated from Stony Brook University with a degree in Computer Science and a specialization in Artificial Intelligence. I joined an EdTech startup from day one and helped scale it to 80,000+ users in under two years, owning and building features end-to-end across the stack, from frontend interfaces to backend systems and infrastructure. I care about writing scalable code, building intuitive interfaces for the best user experience, and solving problems that make a real difference for users.
Languages
When I'm not coding
Basketball
Lifting
Golfing
Beach
Stony Brook University
Bachelor of Science in Computer Science — Artificial Intelligence
Graduated May 2025
Coursework
Courses That Shaped Me
Software Development
Full-stack web development course building progressively more complex applications.
Projects
- ●Reddit clone: iterative build from React/CSS fundamentals to auth, voting, comments, and MongoDB integration
Systems Fundamentals II
Deep systems programming in C covering memory management, concurrency, and low-level data structures.
Projects
- ●Huffman file compressor/decompressor
- ●Custom malloc, calloc, and free implementation
- ●Threaded mailbox system with mutexes
Natural Language Processing
How LLMs work under the hood, from tokenization and embeddings to attention mechanisms.
Projects
- ●Built a custom LLM designed to help students learn abstract studying concepts
Experience
West Palm Test Prep
AI-powered EdTech learning platform serving 80,000+ users, built in a TurboRepo monorepo with Next.js, Expo (React Native), Express, and PostgreSQL via Prisma ORM.
Software Engineer
January 2024 - PresentLaunched whiz.study, an AI-powered EdTech learning platform serving 80,000+ users, built in a TurboRepo monorepo with Next.js, Expo (React Native), Express, and PostgreSQL via Prisma ORM
Integrated a Stripe payment infrastructure including subscription billing, one-time payments, and a free trial system driving 1,600+ active subscriptions and recurring revenue
Led a team of engineers in building an AI tutor using OpenAI SDK and OpenRouter, delivering step-by-step walkthroughs, real-time explanations, and tailored feedback that interacts with the student's workspace to guide them through problems
Refactored a tightly coupled monolith into a modular, plugin-based structure where each exam type is self-contained with its own implementation and architecture, reducing new exam type creation time by ~75%
Expanded the iOS mobile app to support 40+ AP exams, building new React Native views and Express backend endpoints to mirror existing web app functionality
Constructed a concept-level knowledge graph (DAG) to map topic dependencies, powering mastery tracking and personalized study recommendations across the platform
Stony Brook University
Research and engineering roles spanning financial simulations, automated data pipelines, and LLM hallucination reduction.
Software Engineer Intern
September 2023 - December 2023Built an automated web scraping pipeline using Puppeteer to extract current IRS tax bracket data across 4 tax categories, parsing bracket rules across multiple filing statuses and persisting year-stamped records to PostgreSQL
Engineered a Monte Carlo simulation engine with parallel Bun workers to model long-term retirement outcomes, processing 48,000+ financial calculations per run including progressive tax computation, investment growth, and withdrawal optimization
Implemented financial algorithms including Roth conversion tax optimization, Required Minimum Distribution calculations, and custom asset liquidation strategies across 3 account types, enabling users to simulate personalized withdrawal plans
Machine Learning Researcher
May 2023 - August 2023Developed an end-to-end RAG pipeline using OpenAI ada-002 embeddings and Pinecone vector search, reducing GPT-3.5-Turbo hallucination
Benchmarked zero-shot, few-shot, and chain-of-thought prompting across 1,000+ GPT-3.5-Turbo and GPT-4 responses, scoring factual consistency via cosine similarity against ada-002 reference embeddings
Projects
HoopMetrics
A Python desktop application that scrapes live data from nba.com using Selenium. Select any NBA team to view their full roster, coaching staff, and latest team updates. Explore franchise history including retired numbers, Hall of Fame inductees, all-time records, and championships. Compare any two players head-to-head with color-coded stat breakdowns.
Let's Work Together
I'm currently looking for full-time opportunities and open to freelance or contract work. If you have a role, a project, or an idea you'd like to discuss, I'd love to hear from you.