A lab-gallery of shipped systems, AI engines and founder experiments โ each with its own room.
Proof Rooms
A room of trophies. Each flagship proves a major hiring thesis โ advanced AI orchestration, production agents, and live startup execution.


An AI-native writing platform and multi-model orchestration engine โ routing intent across models with memory, canon and evaluation loops.


An AI-assisted local media platform โ lightweight production agent workflows handling editorial automation and content ranking at low cost.

A live local-commerce startup โ a food-rescue marketplace connecting merchants and buyers, with mobile, admin and product all shipped under real friction.
Systems in Orbit
Substantial work reinforcing the story without competing with the flagship rooms.


Leviathan: a quant R&D platform fusing market intelligence (whales, news, sentiment), Strategy-Pattern backtesting and rules-based risk management.


A founder-built e-commerce platform running for real businesses โ catalog, orders and payments owned end to end.


An enterprise AI claims analyzer for insurance: an Intent Router on Azure OpenAI classifies each claim and hands it to a specialized agent for a structured answer.


A full-stack WhatsApp e-commerce agent using a Tool-First Architecture to eradicate hallucinations.


Leviathan: a quant R&D platform fusing market intelligence (whales, news, sentiment), Strategy-Pattern backtesting and rules-based risk management.


A founder-built e-commerce platform running for real businesses โ catalog, orders and payments owned end to end.


An enterprise AI claims analyzer for insurance: an Intent Router on Azure OpenAI classifies each claim and hands it to a specialized agent for a structured answer.


A full-stack WhatsApp e-commerce agent using a Tool-First Architecture to eradicate hallucinations.
Lab Fragments
Prototypes, tools and research fragments. Evidence, not headline acts.


A Python ETL pipeline that turns raw market data into context-enriched JSONL datasets for fine-tuning specialized trading AI models.


A distributed RAG lab split in two: a local 'brain' orchestrating the pipeline and a RunPod GPU 'muscle' serving 4-bit quantized inference.


A RAG-based agentic system performing semantic analysis over large volumes of legal documents.


A GUI tool that consolidates a whole codebase into a single prompt to accelerate AI-assisted development.


A Python ETL pipeline that turns raw market data into context-enriched JSONL datasets for fine-tuning specialized trading AI models.


A distributed RAG lab split in two: a local 'brain' orchestrating the pipeline and a RunPod GPU 'muscle' serving 4-bit quantized inference.


A RAG-based agentic system performing semantic analysis over large volumes of legal documents.


A GUI tool that consolidates a whole codebase into a single prompt to accelerate AI-assisted development.