Senior Product Engineer building end-to-end B2B SaaS and AI-enabled products with TypeScript, React, Python, and FastAPI.
I turn ambiguous business problems into shipped software: product discovery, architecture, frontend and backend implementation, integrations, testing, and delivery.
Over 6+ years in B2B SaaS, I grew from implementation engineering into technical leadership while continuing to build hands-on. My recent work includes workflow automation, LLM evaluation, structured memory, and reliable human-in-the-loop systems.
- TypeScript, React, and Next.js applications
- Python and FastAPI services
- B2B workflows, APIs, and integrations
- PostgreSQL, Supabase, Redis, and background workers
- LLM agents with constrained tools and human review
- evaluation, grounding, and regression testing
- RAG, structured memory, and retrieval pipelines
- prompt versioning and release controls
- product discovery through implementation
- testing, CI/CD, Docker, and observability
- idempotent workflows and explicit failure handling
- collaboration across engineering, QA, DevOps, and business teams
A mobile-first legal CRM for client pipelines, deadlines, notes, documents, and calendar workflows.
| Evidence | Details |
|---|---|
| Status | Public portfolio MVP with a live demo |
| My role | Product design and end-to-end TypeScript/React implementation |
| Tests | 26 Vitest unit tests and 4 Playwright E2E scenarios |
| CI | Not configured yet |
| Known limitations | Demo data only; not intended for real personal or legal data |
Turns Telegram work discussions into reviewed, structured Work Items through a FastAPI API, async Python worker, and human approval flow.
| Evidence | Details |
|---|---|
| Status | Deployed and dogfooded on a VPS; no public bot access |
| My role | Product architecture and end-to-end implementation |
| Stack | Python, FastAPI, PostgreSQL, Redis, Next.js, Docker |
| Tests | 250+ pytest tests across API, bot, worker, and core logic |
| CI | GitHub Actions runs tests and container builds |
| Known limitations | Telegram-only capture and a single reviewer route |
A working evaluation engine for LLM applications with rubric scoring, grounding checks, deterministic safety gates, regression comparison, and human review.
| Evidence | Details |
|---|---|
| Status | Public seeded live demo |
| My role | Product design, evaluation architecture, and implementation |
| Tests | 135 unit tests and 17 runnable evaluation scenarios |
| CI | GitHub Actions runs lint, typecheck, tests, and build |
| Known limitations | Portfolio workbench with seeded data; no claimed external adoption |
A text-first personal workspace for capture, structured memory, retrieval, and AI-assisted reflection.
| Evidence | Details |
|---|---|
| Status | Single-user MVP, dogfooded daily |
| My role | Product design, architecture, and full-stack implementation |
| Tests | Vitest and Playwright suites |
| CI | GitHub Actions currently runs lint and typecheck |
| Known limitations | Open critical issues are documented publicly; some modules remain partial |
A local-first developer tool for prompt versioning, semantic diffs, regression tests, release gates, rollback, and append-only audit history.
| Evidence | Details |
|---|---|
| Status | Working local-first tool; no hosted public service |
| My role | Product architecture and full-stack implementation |
| Tests | 109+ tests |
| CI | GitHub Actions runs lint, typecheck, tests, coverage gates, and build |
| Known limitations | Local workflow and sample data; no claimed external adoption |
6+ years building and delivering B2B SaaS products.
I started as an implementation engineer and grew into technical leadership, coordinating engineers, QA, analysts, DevOps, and business stakeholders while shipping enterprise software. I now combine that delivery experience with hands-on product engineering across TypeScript/React and Python/FastAPI systems.
TypeScript · React · Next.js · Python · FastAPI · Node.js · PostgreSQL · Supabase · Redis · Docker · Azure · Vercel · OpenAI · Anthropic
- Product Engineering and B2B SaaS
- Backend APIs and workflow systems
- AI agents with human oversight
- LLM evaluation and release safety
- Developer experience and automation
These repositories are public engineering case studies: working software, architecture decisions, test evidence, and explicit limitations. I do not label them as community open source unless there is independent usage or contribution evidence.
- reliable agent workflows and tool calling
- evaluation and observability for AI-enabled products
- long-term memory and retrieval
- product patterns for human-in-the-loop automation