SOFTWARE ENGINEER & TECHNICAL FOUNDER

Built with intent.
Made for people.

I'm Andrew. I turn complex problems into intuitive AI products, working closely with the people who use them.

Explore my work View résumé

Currently building at Nomura New York, NY

Lokin, the AI platform I co-founded to help students find their startup team — Lokin · Web platformFitness Tracker AI · iOS

Fitness Tracker AI, my iOS app for photo-based nutrition tracking and personal coaching — Lokin · Web platformFitness Tracker AI · iOS

FROM FIRST PRINCIPLES TO REAL-WORLD PRODUCTSA few things I've built

Ideas are good.
Making them real is better.

2,000+registered users at Lokin

2 weeks 2 daystypical team onboarding at Nomura

0 1products built from the ground up

01 / SELECTED WORK

Built to be used.

From the first conversation to the last mile.
A selection of products I've taken end to end.

LokinPRODUCT EXPERIENCE

01 / 2023–2025

AI PLATFORM · CO-FOUNDER

Lokin

Good people. Great possibilities.

Helping students find their co-founders. I built the product, the AI matching engine, and the feedback loop that brought it to 25 campus organizations.

2,000+registered users800+ peak weekly active users

BEHIND THE BUILD2023–2025

Lokin.

AI-Native Startup Platform

01What I owned

The technical side end to end: the product, the agentic matching layer, the infrastructure, and the weekly iteration loop with users.

02Problem

Match students into co-founder teams with no behavioral data and no ground truth for a “good fit.” Naive LLM matching was slow and sometimes invented reasons two people clicked.

03Approach

An agentic layer over Postgres (profiles) and Pinecone (semantic similarity). An intent gate chose a full LLM call vs. cheap deterministic ranking, with a deterministic fallback so a bad model response never blocked the feed.

04Tradeoff

Capped LLM usage over “always the best match” — slightly weaker tail ranking to hold a 2.0s p95 and predictable cost. Rolled out campus by campus to earn behavioral signal before trusting the model.

05Outcome

2,000+ users, 800+ WAU through 25 campus organizations; the feed stayed responsive even when the model degraded. Honest miss: retention plateaued once novelty faded — recurring reasons to return, not match quality, were the real gap.

06The stack

Next.jsTypeScriptPythonPostgreSQLGemini APILangChainWebSockets

Fitness TrackerMOBILE EXPERIENCE

02 / 2025–2026

IOS APP · SOLO BUILDER

Fitness Tracker AI

Less tracking. More living.

A personal wellness coach that connects the dots. Photo-based meal logging, wearable data, and AI guidance in one native iOS app.

1.2saverage photo-to-meal scan12+ wearable integrations

BEHIND THE BUILD2025–2026

Fitness Tracker AI.

AI Wellness App

01What I owned

Built solo: the Swift app, the computer vision pipeline, the wearable integrations, and the AI coaching layer.

02The product

AI wellness app and personal life coach. A computer vision pipeline logs calories and macros in 1.2 seconds from a photo, wearables sync automatically across 12+ integrations to track sleep and steps, and an AI agent cross-references health data to deliver fitness and lifestyle coaching. Macro estimates land within approximately 10% of manual entries in controlled testing.

03The stack

SwiftGoogle Cloud FunctionsVertex AIFirebaseComputer Vision

Algo PicksMOBILE EXPERIENCE

03 / 2024–2025

MACHINE LEARNING · SOLO BUILDER

Algo Picks

Finding signal in the game.

Live sports data meets ensemble models. A complete pipeline from ingestion and walk-forward validation to calibrated predictions.

4ensemble modelsPython · XGBoost · FastAPI

BEHIND THE BUILD2024–2025

Algo Picks.

ML Prediction Engine

01What I owned

Built solo: the data ingestion, the feature pipeline, model training and validation, and the FastAPI serving layer.

02The product

Real-time sports modeling system. Streams historical and live data through a feature pipeline, trains ensemble models around XGBoost with walk-forward validation, and serves calibrated outcome probabilities over a FastAPI service. Engineering focus: data ingestion, reproducible backtesting, and low-latency inference.

03The stack

PythonXGBoostFastAPISQLTypeScript

Options TraderPRODUCT EXPERIENCE

04 / 2024–2026

QUANT RESEARCH · SOLO BUILDER

Options Trader

A laboratory for better decisions.

Three trading engines, an interactive backtesting simulator, and paper execution through Alpaca. All performance is simulated or backtested.

3strategy enginesResearch & paper trading

BEHIND THE BUILD2024–2026

Options Trader.

Trading Lab

01What I owned

Built solo: all three engines, the Alpaca execution wiring, and the backtesting simulator.

02The product

Quantitative trading lab built around three engines: a delta-hedging service, an LLM-driven playbook selecting option structures across ~50 stocks, and a SPY regime hybrid strategy wired to Alpaca for paper execution. An interactive simulator runs historical backtests across thousands of strategy configurations. All metrics are simulated/backtested.

03The stack

PythonFastAPIPandasNumPyTypeScriptVite

02 / EXPERIENCE

Ownership at
every stage.

Enterprise scale. Startup speed.
The same hands-on approach.

The full résumé

NomuraSoftware EngineerJul 2025 — PresentNew York, NY

Making complex AI and data platforms easier for teams to adopt.

Led or supported integrations that cut typical team onboarding from two weeks to two days.
Built diagnostics and monitoring tools that reduced typical diagnosis time for common issues from about two hours to 30 minutes.

Jun–Aug 2024Software Engineer Intern · Data Fabric / Presto — integrated Presto JDBC and Pulsar SQL.

Jun–Aug 2023Software Engineer Intern · Data Fabric SDK — shipped a production Python SDK.

LokinTechnical Co-FounderAug 2023 — Jul 2025San Francisco, CA

From a blank canvas to a platform used by 2,000+ students.

Owned the product, infrastructure, and agentic matching layer from the first build through campus rollouts.
Simplified onboarding from six steps and 12 fields to three steps and three fields, lifting completion from about 5% to 50%.
Ran 40+ discovery, demo, feedback, and onboarding engagements. Reached 800+ peak weekly active users across 25 campus and club channels.

BASED IN NEW YORK · CURIOUS EVERYWHERE

03 / THE PERSON BEHIND THE PRODUCTS

An engineer's mind.
A founder's instinct.

The best part of building is understanding who you're building for.

I'm drawn to the space where engineering meets real people. I like the discovery conversations, the messy first prototypes, and the moment a complicated system starts to feel simple.

Outside of work, I'm usually playing chess or poker, thinking about markets, surfing when I can, or finding a mountain-bike trail worth getting muddy for.

AI-native productsForward deployed engineeringFrom zero to one

04 / WHAT'S NEXT? OPEN TO THE RIGHT OPPORTUNITY

Let's build
something that matters.

Exploring Forward Deployed & Solutions Engineering
roles in AI and deep tech. Have something in mind?

andrewshatsky1@gmail.com

LinkedIn GitHub

Built with intent.Made for people.

Built to be used.

Lokin

Fitness Tracker AI

Algo Picks

Options Trader

Ownership atevery stage.

An engineer's mind.A founder's instinct.

Let's buildsomething that matters.

Built with intent.
Made for people.

Ownership at
every stage.

An engineer's mind.
A founder's instinct.

Let's build
something that matters.