Full Stack Engineer · Backend & Applied AI
I’m a backend-focused full stack engineer with 5+ years shipping production web apps — well-designed APIs, fast PostgreSQL queries, reliable data at scale. Now applying that foundation to AI engineering: hybrid search, embeddings, RAG — systems built to ship, not just demo.
“Do not judge me by my successes, judge me by how many times I fell down and got back up again.”
Tools of the trade
A grounded stack shaped by production work—shipping clear APIs, dependable data layers, and interfaces that feel effortless.
Also in the toolbox
The journey so far
Five-plus years of owning the full path—from technical decisions and implementation to deployment and optimization.
Worked as a Full Stack Software Engineer at ELO, taking end-to-end ownership of production-grade web applications and backend systems, from technical planning and development through deployment and optimization. I worked across backend, frontend, APIs, databases, search, automation, and media processing, while leading selected projects and providing technical guidance to other developers.
Joined Tuco.co as a Frontend Developer Intern, building responsive React interfaces for an e-commerce platform from UI/UX specifications. I integrated GraphQL APIs with frontend components and developed reusable components that improved UI consistency and simplified feature development.
Selected experiments
A few recent builds spanning retrieval systems, event-driven services, and polished product experiences.
Personal documents are hard to search. Built a local-first RAG system that answers over your own files: hybrid BM25 + semantic retrieval with RRF fusion and cross-encoder reranking, evaluated with Hit Rate@K, MRR, and nDCG@K. Streams grounded answers with inline citations over SSE.
Evaluated retrieval quality and delivered citation-backed answers streamed in real time.
Nextjs@google/genai@huggingface/transformers@lancedb/lancedbminisearchpdf-parse
An event-driven social media backend: independent services communicating over RabbitMQ, backed by MongoDB with Redis caching, JWT + Argon2 auth, and Winston logging — containerized end-to-end with Docker Compose.
Designed independently deployable services with asynchronous messaging, caching, and containerized infrastructure.
Express5MongoDBRedisCloudinaryRabbitMQDocker
Personal finance app for tracking income and expenses: Clerk auth, a PostgreSQL schema for transactions and budgets, and spending analytics with Recharts. Built production-style with TanStack Query, Zod-validated forms, and shadcn/ui.
Built a type-safe finance workflow with validated data, server-state management, and interactive spending analytics.
NextjsReactTypeScriptClerkPostgreSQLPrisma
A trail of small commits
Learning in public, one commit at a time. Browse the recent contribution trail or explore the full graph on GitHub.
Behind the scenes
The small setup behind the code: a focused, Linux-friendly workspace and a few dependable everyday tools.
LinuxUbuntuZorin OS
VS CodeZedWarp Terminal
GitDockerPostmanHoppscotch
VivaldiValentina StudioNeoHtopDiscordNotionFigma
Say hello
Have a bold idea, a knotty problem, or a role that feels like the next chapter? Let’s turn “what if?” into something real.
Have an interesting project, role, or idea? I’d love to hear about it — I’m always open to building, collaborating, and discussing interesting problems.
replies within 24 hoursAsia/Dhaka · UTC+6