Pankaj Saini

Who I Am

AI/ML Researcher Bridging Theory & Real-World Systems

Scientist at CSIR-4PI with expertise in end-to-end AI/ML solutions - from research and modeling to deployment - across HPC infrastructure, QC, healthcare, and intelligent automation.

Applied AI Research

Published researcher in international conferences with contributions spanning predictive analytics, agentic AI, energy-aware computing, and intelligent system monitoring.

HPC & Systems Intelligence

Designing AI-driven optimization frameworks for High-Performance Computing - resource prediction, workload scheduling, power efficiency, and autonomous monitoring using multi-agent architectures.

Healthcare AI

Applying deep learning and NLP to clinical analytics, patient risk stratification, and intelligent decision-support systems - translating medical data into actionable insights.

Agentic AI Systems

Building autonomous multi-agent frameworks for real-time perception, reasoning, and action - enabling intelligent systems that monitor, adapt, and self-heal without manual intervention.

Projects & Publications

Research & Key Projects

Agentic AI-Integrated HPC Monitoring System

Autonomous multi-agent framework for continuous HPC observability - coordinating across power, thermal, workload, and node-health dimensions with self-healing responses and minimal manual intervention.

Predictive Analytics for HPC Resource Management

Forecasting pipelines on high-frequency telemetry streams for resource availability prediction; anomaly detection and pattern recognition to flag inefficiencies and trigger automated remediation.

AI/ML for Healthcare & Clinical Analytics

Predictive ML models for clinical outcomes, patient risk stratification, and health data pattern discovery; deep learning and NLP applied to diagnostic data and decision-support workflows.

NLP & Intelligent Information Systems

Fine-tuned large language models for domain-specific tasks: automated query interfaces, document intelligence, and structured data extraction integrated into operational tools.

Energy-Aware & Green HPC Optimization

ML-driven frameworks for power and thermal management in HPC clusters; research on sustainable, energy-efficient computing strategies for large-scale infrastructure.

Smart Service Monitoring System

Full-stack web portal + Android app for campus-wide service monitoring using deep learning-based image classification, deployed on cloud infrastructure at IIT Delhi.

Technical Expertise

Core Skills

AI / Machine Learning

Deep Learning Time-Series Forecasting Predictive Modeling Anomaly Detection Transfer Learning Reinforcement Learning

NLP & Agentic AI

NLP LLM Fine-tuning Agentic AI Multi-Agent Systems RAG Document Intelligence

Computer Vision

Object Detection Image Classification Real-Time Vision Medical Imaging

Frameworks & Tools

PyTorch TensorFlow Keras Hugging Face scikit-learn OpenCV

Domains

HPC Optimization Healthcare AI Energy-Aware Computing System Intelligence DevOps/MLOps

Tech Stack

Python C/C++ Java AWS Docker OCaml Prolog Agile ELK Stack Git PostgresSQL SQL/NoSQL

Recognitions

Achievements & Background

🏆

2nd Place – WIN Challenge

IIT Hyderabad & WIN Foundation, USA · Image Processing

🥈

2nd Prize – iPreneur'18

Social Venture Challenge, TISS Mumbai

🎖️

Appreciation Award – Red Brick Summit

IIM Ahmedabad · PUZZOKU prototype

🌟

Top 15 Nationally – 3rd NSEIC

Azim Premji University, Bengaluru

💡

Braille Slate Innovation

Sristi-UNICEF Summer School, Prof. Anil Gupta, IIMA

🤝

Robotics Club Mentor

IIT Delhi · Guided teams for competitions