I design and ship production software integrating Large Language Models, sandboxed privacy proxies, and real-time backend pipelines. Practical systems built with Python, TypeScript, Rust, Bun, FastAPI, and PostgreSQL.
Engineered systems covering privacy proxies, concurrent message workers, and relational query engines.
Zero-Retention AI Privacy Sidecar. A local proxy decoupling internal services from third-party LLMs through in-memory token redaction.
WhatsApp automation engine built on Bun with an interactive real-time management console and WebSocket event broker.
Natural language database query layer with dynamic relational schema pruning and AST verification.
Retail intelligence platform featuring normalized relational schemas, multi-stage ETL pipelines, and analytical endpoints.
Core technical pipelines ensuring zero data retention, strict query parsing, and low latency.
Rust core and QuickJS sandbox for in-memory PII scrubbing before LLM egress.
Extracts relevant table schemas and runs sqlglot AST dry-run EXPLAIN checks.
Non-blocking worker lifecycles with RBAC and real-time state synchronization.
Technical leadership, foundation model evaluation, and production data automation.
Core runtimes, languages, and patterns applied in production environments.
| Domain | Toolsets & Runtimes | Architectural Guarantee |
|---|---|---|
| AI & LLM Systems | Gemini API, Claude API, OpenAI SDK, RAG, Embeddings | Deterministic tool calling, schema-bounded outputs |
| Backend Services | FastAPI, Bun, Node.js, WebSockets, PostgreSQL, SQLAlchemy | Sub-80ms analytical latency, non-blocking event loops |
| Privacy & Security | Rust, Embedded QuickJS, sqlglot AST Parser | Zero disk persistence, dry-run EXPLAIN query guards |
| Infrastructure & Tooling | Docker, Linux Server Administration, Git, CI/CD | Containerized reproducible local and production deployment |
Available for AI application engineering, automation systems, and backend infrastructure delivery.