From the system behind the product to the demos and launch content that explain it, I lead the work as one connected build.
I work across the system, product surface, and the story around it—without splitting responsibility between technical and creative teams.
One person.Working stack
Selected engineering evidence
Real systems, concrete contributions, and implementation depth.
A working system connecting GitHub pull requests to LLM drafting, human review, Notion sync, APIs, and persistent state.
My contributionBuilt the automation workflow, backend APIs, database integration, and product interface.
A working system connecting GitHub pull requests to LLM drafting, human review, Notion sync, APIs, and persistent state.
My contributionBuilt the automation workflow, backend APIs, database integration, and product interface.
FastAPI · Next.js · PostgreSQL · Gemini · Notion APIView project sourceA PIGEON-style image-geolocation pipeline using visual embeddings and hierarchical geographic classification.
My contributionUsed frozen CLIP ViT-B/32 features, S2 Geometry, and confidence-based filtering.
Python · CLIP ViT-B/32 · S2 GeometryView project sourceAn audio anti-spoofing system built around an AASIST-style detection pipeline.
My contributionBackend/API and model-integration direction focused on testable inputs and deployable model behavior.
Python · FastAPI · PyTorch · AASIST-style pipelineProject summary availableA testable behavior-profiling workflow covering preprocessing, feature preparation, clustering, and experimentation.
My contributionBuilt the analysis workflow and Streamlit delivery surface with automated tests.
Python · Pandas · Scikit-learn · Streamlit · PytestView project sourceModels, agents, APIs, data flows, evaluation, and dependable automation.
AI engineeringInterfaces and backend behavior shaped around real users and real workflows.
Product deliveryDemos, explainers, launch content, and AI video that communicate what the product changes.
Communication layerAI video is the communication layer of the same build—made to help customers, teams, and audiences understand the product faster.
Turn a script into a credible on-screen presenter without scheduling a studio day.
One shared method keeps engineering, product, and communication decisions aligned.
Clarify the audience, the problem, and what success must look like.
Choose the system architecture or creative treatment before production.
Engineer the working product or produce the complete video asset.
Test, review, revise, and deliver a result ready for real use.
4th Position, All Pakistan Prompt Engineering Competition — 2026
Top 20 Finalist, Global AI Hackathon — 2026
Read the full backgroundSend the problem, audience, constraints, and desired outcome. I will reply with the clearest place to start.
