Mohd Kaif

I’m an AI/ML engineer with hands-on production experience building LLM-powered systems, RAG pipelines, and NLP APIs that handle real workloads. My focus is making retrieval quality, evaluation, and deployment reliability matter in day-to-day systems.

In my Machine Learning Intern role at Acro Technologies, I improved semantic search retrieval relevance by 20%+ using pgvector with cross-encoder reranking over a BM25 baseline. I also reduced manual query routing time by ~60% by converting free-text user queries into structured actions using an LLM pipeline with zero rule-based hardcoding.

On resume understanding, I built a BERT-based NER pipeline that processes 2,000+ resumes, achieving 85–90% entity extraction F1 on domain-specific data. I deployed the system as a FastAPI service on IIS and optimized inference latency by ~40% to support real-time parsing under concurrent client workflows.

Beyond internships, I delivered end-to-end projects such as a Personalized Learning Buddy grounded in document-retrieved context via a RAG pipeline, and a live parking spot detection system using OpenCV-based preprocessing. I’m now targeting AI/ML engineering roles where retrieval quality, system reliability, and production deployment actually matter.