Designs and ships production AI systems that connect large language models to real products, private data, and multi-step workflows.

Application Integration Integrates large language models such as GPT, Claude, and Gemini into user-facing products, including conversational chatbots, semantic search engines, and automated business workflows. Handles API orchestration, streaming responses, tool/function calling, rate limiting, fallback strategies across providers, and cost/latency optimization in production.

System Architecture (RAG) Architects Retrieval-Augmented Generation pipelines that let models securely query private company data. Builds ingestion and chunking pipelines, generates embeddings, and manages vector databases (e.g., Pinecone, Weaviate, Qdrant, pgvector). Implements hybrid search, reranking, metadata filtering, and access controls so responses stay grounded, accurate, and permission-aware.

Prompt & Context Engineering Optimizes how instructions, data, and context are delivered to models to improve accuracy and reduce hallucinations. Designs system prompts, few-shot examples, and structured output schemas; manages context-window budgets; and builds evaluation suites to measure and iterate on output quality, consistency, and reliability.

Orchestration & Multi-Agent Systems Builds complex, stateful AI applications and multi-agent workflows using frameworks such as LangChain and LangGraph. Designs agent roles, tool integrations, memory, routing logic, and human-in-the-loop checkpoints to automate multi-step reasoning and decision-making tasks end to end.

Renaissance-class offensive security researcher with demonstrated global top-tier depth in multiple rare niches and shipped production engineering capability.

I work across cybersecurity, AI/ML, software protection, and blockchain with a focus on problems that demand both depth and breadth. My core work spans reverse engineering and binary analysis across Windows, macOS, Linux, Android, and iOS software protection research on commercial protectors and virtualization systems; mobile application security including native library analysis and runtime instrumentation; and API protocol analysis with authentication internals and anti-bot research.

Beyond security, I build AI-powered products, real-time data pipelines, developer tools, and scalable backend systems. My work spans from native code and assembly-level analysis to full-stack product development, I'm comfortable anywhere on the stack, and most drawn to hard engineering problems where the interesting answers live below the surface.

I'm open to serious collaboration around security consulting, software protection engineering, AI/ML tooling, data infrastructure, and developer automation.

The software I actually have open, and what I build on top of.

• watchman — local-first memory, governance, and code-audit engine for AI agents.

Repositories · Gists · on GitHub since October 2017