Mohd Zahid Faiz
Data Scientist | Applied ML | GenAI & LLM Systems
ZS Associates| since July 2021 |4.9 years
I'm an experienced Data Science professional who works at the intersection of data science, machine learning engineering, GenAI, and LLM-powered systems, building scalable engineering pipelines, data insights, and product solutions to solve complex business problems and deliver measurable impact.
My experience spans leading complex problem statements end-to-end and translating them into practical, data-driven solutions that support decision-making at scale. I've collaborated closely with cross-functional stakeholders to deliver end-to-end data science solutions across consulting and product-driven environments, contributing to revenue growth and improved operational efficiency.
I have hands-on experience designing and deploying Generative AI applications, LLM integrations, and agent-based workflows, enabling intelligent automation and improved user experiences. I enjoy working on problems where data science and AI meaningfully shape outcomes, and I'm always open to connecting with professionals who value thoughtful problem-solving, AI innovation, and real-world impact.
Technical Skills
Languages
Machine Learning / Math
NLP & LLM Systems
Agentic Frameworks
Deep Learning
Data Engineering
Cloud & MLOps
Databases & Tools
Projects
Digital Twin using Clinical GAN
Built a ClinicalGAN-based Digital Twin framework to synthesize high-fidelity patient data. Used multi-stage transfer learning with frozen-layer fine-tuning to stabilize GAN training, evaluated using AID and FID metrics.
Content Authoring using Agentic Framework and RAG
Developed a multi-agent LLM system to auto-generate and refine training content and Q&A from documents. Powered by a RAG pipeline with semantic chunking, cross-encoder reranking, and iterative retrieval for context-accurate generation.
Demand Forecasting for Revenue Management
Architected a PySpark-based distributed forecasting engine over 20B+ data points using key-salting to eliminate skew. Built a market-invariant GBT model via transfer learning and deployed automated MLOps pipelines for price elasticity tracking.
Cold-Start Problem for Newly Launched Drugs
Solved sparse drug adoption data using a PU-Learning Lookalike framework that learns from limited early adopters to score unlabeled HCPs. Output propensity scores directly drove automated sales-force prioritization at market launch.
Graph Network-based KOL Identification
Fused referral networks, claims data, and speaker programs into a unified graph to model HCP influence. Applied PageRank and centrality algorithms for dynamic KOL ranking, driving market share growth from 2.5% to 6% in 12 months.
Churn Prediction and Analysis Platform
Built billion-row PySpark EDA pipelines and Apache Airflow DAGs for end-to-end churn workflow orchestration with PostgreSQL. Achieved 40% deployment efficiency gain through CI/CD automation of production-grade Spark job pipelines.
YOLOv3 Object Detection with Image Classification
Created a two-stage CV pipeline: automated dataset creation via Selenium scraping, followed by YOLOv3 for multi-class object detection coupled with a downstream image classifier. Achieved 88% classification accuracy on the test set.
Education
M.Tech — ECE (Signal Processing & Pattern Recognition)
IIIT Bangalore, India
CGPA: 3.64 / 4.00
- Graduate Teaching Assistant — Machine Learning: Mentored a cohort of 120+ students; conducted technical office hours and evaluated advanced examinations for the Graduate ML curriculum.
- Hackathon Lead: Designed and orchestrated a department-wide ML hackathon; developed the problem statement and automated evaluation metrics for 30+ competing teams.
B.Tech — Electrical & Electronics Engineering
Bhilai Institute of Technology, India
Achievements
Certifications
Generative AI with Large Language Models — Deeplearning.AI
- Built strong understanding of Generative AI lifecycle: data curation, model selection, evaluation, and deployment of LLMs for scalable, production-grade applications.
- Gained in-depth knowledge of Transformer architecture, LLM training, and fine-tuning techniques for domain adaptation and real-world use cases.
- Applied scaling laws, optimization strategies, and inference techniques to balance compute, performance, and deployment efficiency in large-scale AI systems.