Abu Sadat Ansari
I'm Abu Sadat Ansari, a passionate Full Stack Developer and AI Engineer with a knack for building products that are both technically robust and visually compelling. I love working at the intersection of modern web technologies and artificial intelligence.
Career
Work Experience
Software Engineer Intern
SystemStar
- ▹Developed scalable backend modules and APIs using NestJS and also contributed on their own product HeatFleet, an online heating oil marketplace.
- ▹Built a multi-assistant voice AI workflow using VAPI with context management and assistant routing.
- ▹Created an AI video generation pipeline using Redis queues, workers, FFmpeg, and BunnyCDN.
- ▹Built an image-to-transcript-to-audio pipeline that extracted text from images and generated speech output using Hugging Face TTS models to automate audio generation workflows.
- ▹Hosted the Wan 2.1 video model on Runpod GPU infrastructure and optimized resource usage.
- ▹Improved frontend performance through image prioritization, eager loading, and faster rendering.
- ▹Wrote automation scripts for bulk data correction and repetitive operational tasks.
- ▹Integrated ImageKit to serve optimized AVIF/WebP images and reduce bandwidth usage.
Tech Stack
Skills & Technologies
Projects
Things I've Built
Portfolio RAG
This very portfolio website with a RAG-powered AI chatbot that answers recruiter questions about me using my portfolio data as the knowledge base. Built with Next.js, Gemini API, and Framer Motion.
Edemy | LMS Platform
A Learning Management System (LMS) like Udemy built with a React frontend and an Express & MongoDB backend with Clerk Authentication and Stripe Payment integration. Separate Educator account and Student Account Educator dashboard
Education
Academic Background
Bachelor of technology in Computer Science
University of Engineering & Management, Kolkata
- ▹Relevant coursework: Data Structures, Algorithms, Distributed Systems, Operating Systems, Database Management, Computer Networks, Artificial Intelligence
- ▹Capstone project: Siksharthi: A Hybrid Framework for Automated Answer Grading Using BM25 and Semantic Cross-Encoders