$ cat about.txt
About
Backend-focused full stack engineer with 5+ years building APIs and shared platform services — Node.js and NestJS over PostgreSQL, MongoDB and Redis on AWS, with React on the front end for product and internal-facing interfaces.
Core experience is multi-tenant infrastructure serving many clients from one codebase: per-client database isolation with request-time database injection that keeps tenant resolution out of application code, and serverless API layers built to absorb 10,000+ concurrent requests, with long-running work moved off the request path so response times hold under load.
Since January 2026 the work has extended into GenAI in Python: RAG pipelines, LLM API integration, fine-tuning of frontier and open-source models, and agentic systems built on MCP servers with human approval gates — shipped as product features, with the same attention to cost, latency and failure modes as any other backend.
Comfortable owning services end to end and working directly with clients on scope, architecture and delivery.
# APPROACHMost projects don't fail because of the code — they fail because nobody defined the problem properly first. I understand the problem before building it.
What I work on
Multi-tenant SaaS backends
One codebase serving many clients, each with its own database. At Techtiz that meant resolving the client on every request and injecting its database connection through NestJS dependency injection, so application code never has to know which tenant it is serving. Teamstart used a semi-multi-tenant design to give each client customisable scheduling without per-client forks.
Serverless APIs on AWS
API Gateway and Lambda layers sized so capacity follows real demand, with long-running work pushed onto separate functions so the request path stays fast. At TechBucks that layer scaled to 10,000+ concurrent requests; on D[AI]LY it kept LLM generation from blocking the API.
GenAI features in Python
Since January 2026: LLM integration, RAG pipelines, fine-tuning of frontier and open-source models, and agentic workflows on MCP servers with human approval gates — treated as backend features, with budgets for cost and latency and a plan for failure.
Integrations and identity
Third-party services wired in with fallbacks and retries: a ticketing portal at Techtiz, WordPress and Webflow publishing on D[AI]LY, and SAML 2.0 single sign-on across 10+ identity providers at Asian Solutions.
Background
I studied Software Engineering at COMSATS University Islamabad, Lahore Campus, and started working in 2020 while still studying — as a web developer at Asian Solutions, building an enterprise WordPress plugin that implemented SAML 2.0 single sign-on across more than ten identity providers.
From 2022 the work moved to JavaScript backends: serverless APIs on AWS at TechBucks, then a multi-tenant NestJS platform at Techtiz. Since October 2024 I have taken on independent client work, and since August 2025 I have been at 10Pearls, first as a senior backend engineer and now across the full stack of a consumer PDF processing platform.
In 2026 I added GenAI engineering in Python, including the AI Engineer Core Track covering LLM engineering, RAG, QLoRA and agents.
The full timeline is on Experience, and the work itself on Projects.