Shib Kumar Saraf

AI Engineer • Generative AI Architect • Researcher

I am an AI Engineer and Senior Business Analyst with 2.5+ years of experience delivering production-scale AI/ML systems. I specialize in bridging the gap between research innovation and real-world deployment, with a focus on Gen AI, Agentic Workflows, RAG Architectures, and Computer Vision.

Currently, I drive data-driven strategies at EXL Service and previously led AI engineering initiatives at C-DAC, serving over 10,000 users.

Oct 2025 – Present

• Impact Analytics: Conducting actuarial factor analysis for CVS Health (Aetna) insurance products to support pricing strategies.
• Retention Strategy: Engineered churn analytics models that improved customer retention from 49% to 60%.
• Stakeholder Enablement: Delivering actionable insights to plan sponsors to optimize insurance product performance.

Jul 2024 – Oct 2025

• Scalable Chatbots: Architected ChatDB and FileQAExpert, serving 10,000+ monthly queries via Web, Telegram, and WhatsApp using RAG pipelines.
• Anomaly Detection: Deployed real-time monitoring systems for government exams, improving efficiency by 40%.
• Process Automation: Reduced manual oversight by 60% through automated financial workflow applications.

Jan 2024 – Jun 2024

• Social Sentinel: Developed a real-time NLP system for monitoring and classifying offensive content on Telegram.
• Research: Contributed to IEEE publications through applied experimentation with transformer-based classification.

• Best Paper Award (IEEE ICRITO 2025) Received for the paper "Real-Time Vigilance: Advanced Video Analysis for Identifying Violence in Public Spaces".
• Best Project Award (IIT Madras) Awarded for excellence in the Modern Application Development course.
• Top 3 Rank Holder (MCA) Secured top 3 position in the Master of Computer Applications program at Pondicherry University with a CGPA of 9.35.
• Swami Vivekananda Merit-cum-Means Scholarship Awarded by the Government of West Bengal for consistent academic excellence (2015–2022).

Conference	Paper & Impact
	Real-Time Vigilance: Advanced Video Analysis for Violence Detection 🏆 Best Paper Award at ICRITO 2025 Built a real-time computer vision pipeline leveraging deep learning models for automated violence detection to support public safety monitoring systems.
	Hybrid QLSTM Framework for Time-Series Forecasting Presented at ICIMIA 2025 Designed a hybrid Quantum-inspired LSTM architecture to improve forecasting accuracy under highly volatile financial market conditions.
	Multifaceted Feature Extraction for Mortality Risk Prediction Presented at ICSTSDG 2024 Developed an ML-driven clinical risk prediction framework using advanced feature engineering to improve mortality assessment in pediatric intensive care units.

• Architecture: FastAPI-based agent system with role-scoped isolation.
• Core Logic: Intelligent routing between SQL databases and Vector Stores for grounded RAG responses.
• Features: Schema-aware natural language querying and multi-turn reasoning.

• Tech: LLM-driven NLP pipelines for sentiment modeling and thematic clustering.
• Outcome: Resilient inference pipelines to identify latent behavioral insights from social streams.

• Algorithms: Collaborative Filtering, ALS, and Graph-based similarity modeling.
• Innovation: Hybrid recommendation system using lyrical semantics and audio feature embeddings.

• Master of Computer Applications (MCA) | Pondicherry University | CGPA: 9.35/10
• Diploma in Data Science | IIT Madras | CGPA: 8.8/10
• B.Sc. Physics (Honours) | Vidyasagar University | CGPA: 9.08/10

• Google Digital Leader | Coursera (2024)
• Artificial Intelligence for Economics | IIT Kharagpur (NPTEL, 2025)
• Deep Learning for Computer Vision | IIT Hyderabad (NPTEL)
• Data Science & ML using Python | NIELIT
• IoT Software Analyst | NIELIT
• Introduction to MongoDB | MongoDB University
• Advanced Excel Training | Udemy

"Engineering Intelligence. Deploying Impact."

© 2026 Shib Kumar Saraf