Henry D.

I build production-grade AI software, ML systems, backend infrastructure, and data products. My work spans the path from data acquisition and ML pipelines to LLM applications, agentic workflows, production APIs, developer infrastructure, and user-facing products. I work across the stack when the product requires it, from architecture and data models through backend services, AI/ML systems, frontend applications, testing, security, observability, and deployment. AI APPLICATIONS & AGENTIC SYSTEMS • LLM applications, RAG, semantic search, embeddings, reranking, and agentic workflows • LangGraph multi-agent systems with supervisor and specialized sub-agent architectures • Retrieval using pgvector, FAISS, BM25, hybrid search, RRF, and cross-encoder reranking • AI evaluation, structured outputs, tool calling, document intelligence, and workflow automation • OpenAI, Claude, Groq, Llama-family models, Whisper, and other production AI integrations ML & PREDICTIVE SYSTEMS • Credit risk, fraud detection, recommendation systems, customer analytics, and forecasting • XGBoost, LightGBM, scikit-learn, PyTorch, Prophet, and time-series workflows • End-to-end ML lifecycle work: data preparation, feature engineering, training, evaluation, serving, monitoring, and iteration • FastAPI model-serving systems and production data pipelines BACKEND & API ENGINEERING • Python, TypeScript/JavaScript, Java, and Go • FastAPI, Express, Hono, tRPC, gRPC, REST, and event-driven services • PostgreSQL, Firestore, Supabase/Postgres, Redis, Kafka, BigQuery, and vector databases • Authentication, authorization, API security, rate limiting, idempotency, background jobs, webhooks, and observability • Docker, CI/CD, cloud deployment, Prometheus, and Grafana DATA ACQUISITION & INFRASTRUCTURE I build systems that collect, process, validate, and serve data at scale. My Go-based Web Harvester is a distributed crawling and data-acquisition system with concurrent workers, Redis Streams, PostgreSQL, headless Chromium, proxy health tracking, adaptive rate limiting, retry/backoff, scheduled harvesting, an HTTP job API, Prometheus/Grafana monitoring, structured extraction, and integration/e2e testing. It handles metadata, headings, OpenGraph, JSON-LD, visible text, links, custom selectors, robots.txt, rate limits, blocking, and selector drift. DEVELOPER INFRASTRUCTURE & SAAS I also build developer-facing infrastructure and SaaS products. Schema-Watch monitors API contracts and schema changes, with local-first workflows, CLI tooling, CI checks, affected-file detection, SARIF output, authentication, API keys, RLS, rate limiting, metering, and a path toward hosted team monitoring and integrations. FULL-STACK PRODUCT ENGINEERING When the product needs a complete application, I can build the frontend as well as the underlying systems. My experience includes React, React Native, Expo, Next.js, Tailwind, Zustand, React Query, Firebase, Supabase, and PostgreSQL. I have built full-stack marketplace systems with mobile and web applications, backend services, serverless functions, subscription/payment integrations, role-based administration, chat, notifications, security rules, testing, and deployment workflows. SELECTED WORK • Nigerian fintech fraud intelligence platform combining transaction signals, regulatory retrieval, agentic reasoning, audit logging, and a production API. • Multi-agent deal intelligence system for research across filings, exchanges, and news. • Telecom subscriber analytics and forecasting systems with automated data refresh and API serving. • RAG/document applications using pgvector, embeddings, semantic search, and document editing. • Python gRPC backend refactoring focused on performance, security, and maintainability. • Compliance and operational systems using Supabase/Postgres, RLS, scheduled workflows, and AI-assisted explanations. HOW I WORK I focus on building systems that can actually run, not just prototypes. I care about architecture, clean interfaces, reliable data flows, security, testing, observability, deployment, and maintainability. I can join at different stages: designing a system from scratch, taking an AI prototype toward production, extending an existing backend, building data infrastructure, integrating AI into an existing product, or owning a feature from architecture through deployment. STACK Python • TypeScript • JavaScript • Go • Java • SQL FastAPI • Node.js • Express • Hono • gRPC • REST • GraphQL React • React Native • Expo • Next.js PostgreSQL • Firestore • Supabase • Redis • Kafka • BigQuery LangGraph • LangChain • pgvector • FAISS • BM25 • Reranking XGBoost • LightGBM • PyTorch • scikit-learn • Prophet Docker • AWS • Azure • GCP • Firebase • GitHub Actions Prometheus • Grafana • dbt • Prefect • PySpark I’m available for freelance, contract, and full-time engagements involving AI applications, ML systems, backend engineering, data infrastructure, and full-stack product development.