Omar Farooq Bandial

I’m a Junior AI/ML Engineer focused on agentic AI, LLM applications, and RAG systems, and I love turning ideas into end-to-end working products. My work blends NLP modeling, practical application logic, and production-minded deployment choices.

In my internships, I shipped NLP functionality and built evaluation-driven pipelines. At TenX, I delivered NLP question generation for edtech by automating quiz creation from raw educational content, and I evaluated T5, BART, and distractor strategies to pick the strongest pipeline using ROUGE benchmarks. At Arcana Info, I built a TF-IDF/Logistic Regression ticket classifier on 10K+ tickets with 85% department/priority accuracy, then deployed a Flask REST API using NLTK/spaCy preprocessing for real-time ticket routing.

My key project, AI-Powered Interviewer (FYP), is a serverless AI recruitment platform featuring resume parsing, personalized questions, async interviews, and LLM evaluation. I implemented hybrid resume parsing (pdfplumber with Gemini 2.5-Flash vision fallback), built an “LLM-as-a-judge” evaluation engine with evidence-first scoring, integrated Whisper STT and OpenAI TTS with anti-cheat logging, and deployed it with AWS Lambda/S3/SQS/Step Functions plus PostgreSQL (RDS) and DynamoDB for persistence and metadata.