Senior full-stack & AI engineer. I help founders and engineering teams ship scalable AI agents, RAG systems, and full-stack apps that actually survive contact with real users.
Replace repetitive ops work with agents that just work — multi-step, tool-using systems built on LangChain, CrewAI, or LlamaIndex on top of OpenAI / Claude. Not chatbots. Workers.
Ship a product with AI baked in, not bolted on — auth, billing, dashboards, vector search, and real-time UX. Next.js or MERN on the front, FastAPI or Node behind it.
Ground your LLM in your own data so answers are actually trustworthy — chunking strategy, embeddings, hybrid search, eval harnesses. The part that decides whether your AI feature ships or gets pulled.
Cut the repetitive 60-80% of a workflow, not the headcount that's actually needed. Concrete deliverable: a written plan with hours saved and dollars cut, then I build it.
Problem. Support team of 12 was drowning in repetitive Tier-1 tickets. Founders wanted to scale users 3× without growing the team.
Problem. Compliance team needed to answer policy questions against 4M+ regulatory PDFs without leaking data to public LLMs.
Problem. Patient onboarding took 47 minutes of clinical staff time per case. Repetitive form-filling, EHR lookup, eligibility checks.
Problem. Legacy PHP storefront was buckling at 50K MAU. Needed a rebuild — and the founders wanted AI-driven product discovery, not just a port.
Problem. Building production bots meant stitching together channel logic, workflow execution, AI tools, integrations, credentials, and versioning across disconnected systems.
Pick a time in your timezone. No sales pitch—just a practical product discussion.
† Client names redacted under NDA. Reference calls available on request after a discovery call.
We figure out if I can actually help. If I can't, I'll point you to someone who can. No deck, no sales pitch.
Written proposal: problem, approach, stack, milestones, fixed price or T&M. You sign or you don't.
Daily commits, weekly demos, shared Linear board. You see progress every day. No black boxes, no surprises.
Production deploy, observability, runbooks, handoff session with your team. 30 days of free bug-fix support after.
I'm a senior full-stack & AI engineer with 7 years of production experience. I started as a MERN dev, went deep on Python and ML, and now spend most of my time at the intersection — turning AI prototypes that barely work in a notebook into systems that survive real users, real load, and real money.
I work best with founders who want one senior generalist instead of three juniors, and with engineering managers who need a surgical contractor — not someone they have to onboard for a month.
I won't take a project I don't think will succeed. If your idea needs a different shape, or doesn't need AI at all, I'll tell you on the discovery call. Wasted time is more expensive than an honest no.
In 8th grade, a classmate told me his brother was a "hacker" — and I decided I wanted to be one too. The catch: we didn't have Wi-Fi at home, so I'd borrow a signal from a friend's house just to get online. I earned good grades so my father would buy me a secondhand laptop — 2GB of RAM, 80GB of storage — and tried to run Kali Linux on it. It hung. It crashed. It never worked.
That failure turned out to be the best thing that happened to me. Instead of breaking into networks, I got curious about how they — and the software running on them — actually worked. I picked up Python in 9th grade, built my first small programs, and taught myself HTML, CSS, and JavaScript after realizing "blogging tools" weren't real development. By 10th grade I was shipping small full-stack projects on my own — no bootcamp, just documentation, trial, and a lot of broken builds.
That's still how I approach engineering today: when the direct path is blocked, find the more durable one. Now I make architecture decisions alongside CTOs and build AI agent systems that take real workload off real teams. The most "critical problem" I try to solve for a client is rarely the flashiest one — it's usually the boring, repetitive bottleneck nobody else wanted to fix.
None of this happens without my parents backing a laptop purchase on a promise of good grades, the friend whose Wi-Fi — and stories — started it all, and my siblings who never stopped believing this would go somewhere. — Taha
For a defined feature, MVP, or migration with clear boundaries.
Ongoing senior bandwidth — your roadmap, my keyboard.
I review your stack, workflow, and AI surface area. You get a written plan.
† Prices are starting points. Larger or more complex engagements are quoted after a discovery call.
Book a call directly, or send the details first and I’ll reply within one business day.
Book a 30-min discovery call →