Meta
Core developer for Meta AI and wearables identity infrastructure across Meta AI, Quest, and Ray-Ban smart glasses. Improved registration and account recovery flows.
Distributed systems / AI infrastructure / identity platforms
Jack Su
I build high-scale systems where product growth, secure identity, ML/AI infrastructure, and practical intelligence features have to work together under real traffic.
Senior Software Engineering Leader
Identity: Meta AI, Quest, Ray-Ban smart glasses, SSO, OTP recovery.
Scale: Amazon Photos, distributed data pipelines, high-availability release systems.
AI/ML systems: agentic AI, RAG, ML inference, data pipelines, privacy-first infrastructure.
100M+
users supported across identity and cloud services
100K
requests per second peak infrastructure experience
$50B
monthly equity volume in trading analytics systems
3
AI competition tracks across Kaggle reasoning and game agents
Work highlights
Systems that scale, recover, and earn trust.
Core signal
A career spent where consumer-scale products, secure identity, and dependable infrastructure all have to move together.
Amazon
Built Amazon Photos and Cloud Drive services, distributed data pipelines, facial recognition integration, and release systems with 99.99% availability targets.
Deutsche Bank
Modernized global equity trading and analytics systems, including a 10 TB analytics platform integrating 200+ upstream systems.
Fidelity, SAS, Aether, S&P
Built high-availability authentication, payroll, workflow, emergency-response, and real-time financial market data systems.
GitHub highlights
Recent public builds and experiments.
TypeScriptGitHub
TrendLens
Trend intelligence dashboard for spotting rising topics using public data and AI.
JavaScriptGitHub
X Bookmark Summarizer
Chrome extension package and launch kit for summarizing saved X bookmarks.
TypeScriptGitHub
requestsolution
Community product ideation platform where AI turns ideas into buildable specs.
TypeScriptGitHub
learn-claude-code
A compact coding-agent harness exploring how far shell-first automation can go.
JavaScriptGitHub
market-expert-mcp
MCP experiment for market expert workflows and model-connected financial tooling.
PythonGitHub
TradingAgents
Multi-agent LLM financial trading framework exploration.
Kaggle competition work
Reasoning benchmarks and strategic game agents.
My Kaggle work sits at the edge of efficient reasoning systems: ARC-style visual transformations, minimal neural networks, and game-playing agents that must make strong decisions under simulation constraints.
NeuroGolf 2026Kaggle
Minimal ONNX networks
Profiling candidate networks, repairing static-valid models, comparing them against anchors, and finding task-level cost improvements for ARC-AGI-style transformations.
- Model-size and cost profiling
- Anchor comparison workflow
- Task-level delta ranking
ARC Prize 2026Kaggle
ARC-AGI-2 solver systems
Building around the ARC-AGI-2 static reasoning benchmark: local task analysis, candidate generation, validation loops, and efficiency-aware submission packaging.
- Abstract grid reasoning
- Program-search style iteration
- Private-eval discipline
Orbit WarsKaggle
Orbital strategy agents
Treating Orbit Wars as an agent engineering problem: simulate planet and fleet state, rank attack and defense options, test against baselines, and iterate ladder-ready bots.
- Forward simulation
- Heuristic policy tuning
- Multi-agent strategy testing
Resume
Senior engineering leadership with hands-on depth.
The resume covers distributed systems, identity, AI/ML integration, privacy-first engineering, high-volume finance, and cloud platform work across consumer and enterprise environments.
Architecture: microservices, Kubernetes, AWS, agentic AI, RAG, vector DBs.
AI/ML: LLMs, CNNs, predictive analytics, NLP recommendation engines.
Core stack: Java, Python, TypeScript, JavaScript, Spring, performance profiling.
Personal AI chat