I've worked on production websites and web apps used by real businesses, including a B2B corporate site for an apparel sourcing company and an online store for a fashion brand.
My focus is software that takes repetitive work off people's plates. I've built a CRM that scores and follows up on leads, a receptionist that books appointments over chat and voice, and an assistant that answers a business's Facebook, Instagram and WhatsApp messages around the clock.
What I bring
- Ownership: I take a feature from idea to production: UI, backend, AI integration and deployment
- Clarity: Specifications, acceptance criteria and ADRs before code; scope and trade-offs documented in the repo
- Quality: Tested, secure and observable code that holds up in production, not just in a demo
- Collaboration: Small, focused PRs backed by code review and CI, equally effective in a team or solo
|
Software Engineer 🏢 Brain Stack | London · New York · Dhaka | Remote (Contract) • Build and ship client web apps with Next.js, TypeScript and Node.js, from scoping to deployment • Integrate AI features such as LLM chat, booking agents and workflow automation into client products • Maintain production sites for performance, SEO and reliability, and review teammates' code |
|
Full Stack Developer 🏢 Digiwinners | Uttara, Dhaka · Remote | Full-time • Build responsive, scalable and secure web applications end to end • Develop REST APIs, authentication and database schemas behind clean, reusable UIs • Ship features to production through Git workflows, pull requests and code reviews |
|
Full Stack Developer Intern (Frontend Focused) 🏢 Digiwinners | Uttara, Dhaka · Remote • Built responsive, high-performance web apps with Next.js, React and TypeScript • Implemented state management and API integration for data-driven pages • Worked in a team Git workflow with feature branches, pull requests and reviews |
| Project | What it does | Stack | Links |
|---|---|---|---|
| AI CRM & Sales Automation | Multi-tenant CRM with AI lead scoring, email drafts and an automation engine with human approval; 268 tests, CI. | NestJS · Next.js · PostgreSQL (RLS) · Prisma · Redis/BullMQ · Anthropic | Live · Code |
| AI Health Receptionist | Books clinic appointments over chat & voice: LLM tool-calling, Google Calendar sync, switchable AI providers, Twilio calls. | NestJS · Next.js · PostgreSQL · Twilio · Anthropic/Gemini/OpenAI/Groq | Live · Code |
| Social Ops AI Automation | Runs a business's Facebook, Instagram & WhatsApp: daily posts, RAG auto-replies, comment moderation, lead extraction, urgency alerts. | Express · TypeScript · MongoDB Vector Search · Gemini · React | Code |
| Faith Journey | Modest-wear e-commerce for women (frontend, team): 20+ type-safe components, query caching cutting redundant API calls ~15%. | Next.js · TypeScript · TanStack Query · React Hook Form · Zod · shadcn/ui · Tailwind CSS | Live |
| Cloths 'R' Us | Corporate B2B site for an apparel sourcing company: 9 SEO-optimized pages, product catalog, images cut from 335 MB to 32 MB. | Next.js 16 · React 19 · TypeScript · Tailwind v4 · GSAP · Netlify | Live |
| Category | Skills |
|---|---|
| Languages | TypeScript, JavaScript, Python, Dart, C/C++, PHP, SQL |
| Backend | Node.js, NestJS, Express, Fastify, FastAPI, REST APIs, Socket.IO, Laravel |
| Frontend | Next.js, React, Tailwind CSS, Redux Toolkit, Zustand, TanStack Query, GSAP, Framer Motion, Three.js |
| Mobile | React Native, Expo, Flutter |
| Database | PostgreSQL, Prisma, TypeORM, MongoDB (Vector Search), Redis |
| AI | Anthropic, Gemini, OpenAI, Groq, RAG, embeddings, tool calling, structured output |
| Infra & Quality | Docker, GitHub Actions, BullMQ, Jest, Vitest, Playwright, Vercel, Railway, Render |
| Integrations | Stripe, Twilio, Meta Graph API, WhatsApp Cloud API, Google Calendar/Sheets, Cloudinary, Clerk |
🔭 Currently building: AI voice agents & workflow automation · 📚 Learning: AI agents, MCP and LLM evaluation, tracked in ai-engineering-journey
- Plan before building: SRS, milestone plan and ADRs live in the repo; the AI CRM shipped all 9 planned milestones against them
- Test against reality: Integration tests on real Postgres and Redis, Playwright E2E for user journeys, CI on every push
- Secure by default: Postgres RLS tenant isolation, RBAC, Zod validation, rate limiting and HMAC-verified webhooks
- Design for failure: Idempotent jobs, retries with backoff, crash-safe queues, real health checks and request-ID logs
- AI with guardrails: Provider-agnostic LLMs, deterministic test stubs, structured outputs and human approval on customer actions
