# SJ Cha

**Product Lead, AI Systems** — San Francisco, CA, United States

> AI product manager with 8 years building applied AI features (ASR, TTS, LLM, agentic systems) at frontier companies, from 0→1 prototyping through scaling to millions of users.

Career Archetype: **Launchpad Navigator** (HS-DG) — One of the first Launchpad Navigators on Saywise

## Links

- LinkedIn: https://linkedin.com/in/sj-cha

## Experience

### Product Lead, AI Systems (Special Projects), Speak (2024-12 – present)
San Francisco, California, United States · Prototyped agentic curricular design system that cut production time from 8 to 2 weeks (75% reduction). Building v2 with MCP servers and knowledge graphs to scale across language pairs and proficiency levels. Established evaluation frameworks using Braintrust, SQL, and Python to measure content performance, and shipped new voice-based lesson experience using OpenAI Realtime API.

### Product Manager, Learning Experience & Infrastructure, Speak (2023-08 – 2024-11)
San Francisco, California, United States · Productionized frontier AI with pre-release GPT-4 access, rapidly prototyping conversational AI features (Free Talk, AI Q&A) that set industry design patterns. Updated in-app TTS with more natural delivery via OpenAI RT Audio and Gemini RT Stream. Automated mass lesson content generation via prompt chaining, reducing what took months to three days.

### VP of Operations / Head of International Expansion, Speak (2021-07 – 2022-12)
San Francisco, California, United States · Invented the Speak Method framework for language learning, establishing initial product-market fit. Scaled mobile-first language learning from 3-person pre-product startup to unicorn valuation with $100M+ ARR. Led cross-functional execution across product, growth, marketing, and operations through 0→1→n scaling phases.

### Founding Member / Head of Operations, Speak (2018-01 – 2020-01)
San Francisco, California, United States · 0 -> 1 – Invented the "Speak Method" framework of language learning and defined the core UI/UX via content and product that established early PMF, later scaling mobile-first language learning experience from a 3-person pre-product startup to a unicorn

### Product Analyst & Data Scientist, Rent The Runway (2015-06 – 2017-12)
New York, New York, United States · Prototyped a personalized recommender engine to increase subscriber engagement and retention. Analyzed growth levers and shared actionable insights to reduce churn.

## Education

### B.A. with Honors in Applied Mathematics & Statistics, Harvard University
1st place in a predictive modeling competition for a graduate-level Generalized Linear Models course

## Projects

### Paryn
Personalized parenting guidance service · Rapidly prototyped full-stack mobile app using Claude Code to explore personalized agentic chat with memory, validating product concepts while working as a non-traditional engineer.

### Reservation Voice Agent
Voice-based restaurant reservation system · Built a voice agent to make a reservation at Saison using Lindy.ai, demonstrating practical application of voice AI for real-world scenarios.

### Internal Audio QA Validation Tool
Automated audio stream validation from RT audio model · Prototyped an internal tool to auto-validate and capture audio streams from RT audio models. The ML team later productionized it, achieving 80%+ efficiency gain in Tutor Lesson audio QA.

## Skills

SQL, Tableau, Product Analysis, Statistical Data Analysis, MCP, Product Management, Multi-agent Systems, Large Language Model Operations (LLMOps), Data Analysis, User Research, Claude Code, PromptLayer, Braintrust, Cursor, Whisper, Azure TTS, ElevenLabs, Minimax, Google AI Studio, Magic Patterns, Make, Syncly, Devin, Indent, Notion AI, Veo, Sora, Mode, Retool, n8n, TypeScript, Node.js, Python, GCS, Prompt Engineering, Context Engineering, Evaluation Framework Design, Agentic System Architecture, ASR, TTS, LLM, S2S

---
Source: https://saywise.com/member4016 (last modified 2026-10-05T21:09:18.224Z)