Professional Experience
AI/ML Engineer Intern - Gurugram
Quess Corp
- Shipped an AI voice telecaller system (Sarvam AI + Groq Qwen-3) handling 500+ daily candidate interviews, cutting manual screening time by 60% and saving ~4 hours/day of recruiter effort.
- Built a production intent classification chatbot using SetFit/FastFit few-shot learning, achieving 94%+ accuracy while fully eliminating LLM hallucination risk via deterministic classification.
AI/ML Engineer Intern - Remote
IBM
- Architected a Multi-Output Gradient Boosting system to predict personalized treatment plans simultaneously, utilizing synthetic data generation to resolve class imbalance and improve recall by 15%.
- Developed interactive Streamlit dashboard integrating NHANES pipelines, reducing patient risk profiling time by 40%.
AI/ML Intern - Delhi
EY
- Deployed a domain-specific Chatbot by Fine-tuning open-source LLMs, reducing query resolution time from 12 to 9.6 minutes (20% improvement) for 200+ daily users.
- Constructed automated ETL Pipelines to harvest and clean data, creating a robust corpus that reduced hallucination by 25%.
Research Intern - Remote
NIT Kurukshetra
- Built a Neural Architecture Search (NAS) engine using Differential Evolution in PyTorch, automating CNN topology discovery and improving accuracy by 5% on CIFAR-10.
- Coded dynamic probability adaptation algorithms for genetic operators, accelerating search convergence by 10%.