01Sole architect & lead engineer
Book Generator AI
An AI-native publishing platform built around the hardest problem in long-form text: keeping context coherent across tens of thousands of words. A multi-stage pipeline (planning, chapter generation with context propagation, and an automated evaluation engine) turns a rough idea into a publication-ready book that reads in one voice from first page to last.
stack
Next.js · NestJS · Postgres · Redis
ai layer
LLM orchestration · multi-agent · eval
scope
architecture → payments → on-call
outcome
long-form context solved · coherent 40k-word books
$ deploy book-generator-ai --env prod ✓ real books live on Amazon KDP
02Full-stack + AI engineer
Enterprise knowledge & RAG
Retrieval-augmented systems over private corpora: document ingestion, chunking, embeddings and vector search, serving grounded answers through LLM-powered interfaces that internal teams actually use.
stack
Python · TS · vector DB · Postgres
ai layer
RAG · embeddings · tool calling
scope
pipelines + grounded UX
outcome
grounded answers · eval-gated
$ rag eval --suite regression ✓ relevance up, hallucination down
03Backend & platform engineer
Cloud-native platforms & APIs
Distributed backends and event-driven architectures: REST and microservices, auth, queues, caching and observability, built to serve thousands of users and stay deliberately boring under load.
stack
Node · NestJS · Go · AWS · Docker
pattern
event-driven · queues · cache
scope
APIs · auth · observability
outcome
thousands of users · zero-downtime deploys
$ kubectl rollout status ✓ zero-downtime
04Full-stack mobile engineer
Mobile & cross-platform apps
Cross-platform products in React Native with offline-first thinking, clean state management and a codebase that stays maintainable as the roadmap keeps growing.
stack
React Native · TS · REST
pattern
offline-first · clean state
scope
iOS + Android, one codebase
outcome
shipped to stores · easy to extend
$ expo build --platform all ✓ shipped to stores
