Forging Autonomous AI Models & Sub-Millisecond Full-Stack Cloud Engines.
Staff-level Full-Stack & AI Systems Engineer with 6+ years shipping production architectures. Proven record deploying multi-agent LLM harnesses handling 42M+ tokens/day and consumer-scale web platforms serving 12M+ MAUs with sub-36ms edge latency.
Live Interactive Engineering Sandboxes
Real code, real algorithms, and real telemetry running live right in this browser session.
LLM Inference Cost & Latency Arbitrage Simulator
Simulate your production token volume. Compare standard unoptimized API costs against the speculative caching & semantic compression engine I designed.
By combining AST token prefix matching, speculative lookahead decoding on cheap edge models, and KV-cache recycling, prompt ingestion overhead is cut by up to 68%.
Flagship Systems Architecture
Explore the system diagrams, engineering trade-offs, and verifiable metrics behind my key projects.
NexusAgent — Autonomous Multi-Agent Orchestrator
Designed and deployed a production multi-agent harness powering enterprise code generation and data analysis. Solved the token bloat and multi-step latency compounding problem through speculative lookahead decoding and sliding-window KV-cache recycling.
Verified Technical Matrix
Languages, frameworks, and infrastructure tools battle-tested in high-stakes production.
- PyTorch & HuggingFaceAdvanced
- vLLM & Speculative DecodeProduction
- LangGraph & Agent LoopsProduction
- Qdrant & Hybrid RRFAdvanced
- ColPali & Multimodal RAGProduction
- Next.js 16 (App Router)Expert
- React 19 & Server ActionsExpert
- TypeScript (Strict Mode)Expert
- Tailwind CSS 4 & ShadcnExpert
- OffscreenCanvas & WorkersAdvanced
- Python (FastAPI/AsyncIO)Advanced
- Rust & WebAssemblyIntermediate
- PostgreSQL & pgvectorExpert
- Redis Clusters & StreamsAdvanced
- WebSockets / Binary CRDTAdvanced
- Cloudflare Workers & DOAdvanced
- AWS (ECS, Lambda, S3)Production
- Docker & K8s DeploymentProduction
- OpenTelemetry & DatadogProduction
- CI/CD & Automated TestingExpert
Engineering Track Record
Roles at high-growth tech companies delivering measurable, revenue-driving systems.
Staff Founding Engineer @ HyperScale AI
Architected the core distributed agent runtime processing 42M+ daily tokens. Led a squad of 6 engineers across frontend and ML infrastructure. Re-engineered model routing logic, dropping inference cloud spend by 68% and maintaining 99.99% uptime through two major product launches.
Senior Full-Stack & Systems Engineer @ FinPulse Systems
Spearheaded the real-time financial market analytics terminal handling 250,000 WebSocket events/second. Eliminated UI frame drops by offloading calculations to Web Workers with SharedArrayBuffer. Modernized the legacy dashboard to Next.js with React concurrent features, boosting user engagement by 43%.
Core Contributor • AI Agent & Edge Tooling
Created and maintain widely adopted developer utilities for LLM prompt caching, token serialization, and lightweight state reconciliation in Next.js applications.
Zero-Friction Hiring & Direct Channel
Everything engineering leaders and talent recruiters need to evaluate, screen, and extend offers without back-and-forth delays.