Headline

AI/ML Engineer @ Jio Platforms | Generative AI | RAG Pipelines | Multi-Agent Systems and Workflows | LLM Fine-Tuning and Deployment

About

As AI ML Engineer at Jio Platforms Limited, I specialize in developing AI-driven solutions for real-world challenges. My contributions include creating a RAG-based government AI chatbot to streamline public policy queries, optimizing retrieval strategies to reduce hallucinations by 55%, and building automated pipelines to enhance data processing efficiency. These efforts have significantly minimized OpenAI API calls and improved the chatbot’s accuracy. With a degree in Artificial Intelligence from G H Raisoni Institute of Engineering and Technology, I bring expertise in data science and core technologies like NLP, multimodal models, and computer vision. By leveraging my skills in SQL, PostgreSQL, and Python, I aim to drive innovation and scalable solutions in the AI domain, transforming complex challenges into impactful technological advancements.

Location

India

AI stack

antigravity

claude code

cursor

hugging face

langchain

langgraph

Work experience

AI/ML Engineer at Jio Platforms Limited (JPL), Jan 2024 - Present · 2 yrs 8 mos, On-site, Building and deploying production-grade AI systems serving 1M+ users across Jio's consumer and enterprise applications. Focused on Generative AI, multi-agent systems, RAG pipelines, computer vision, and LLM deployment. 🔹 Built a RAG-based Government AI Chatbot for Indian government public policies. Reduced hallucinations by 55% through retrieval and chunking optimization, and cut OpenAI API costs by 57% via caching and guardrails. 🔹 Engineered a production-grade multi-agent content moderation pipeline using LangGraph and CrewAI. Improved accuracy from 68% to 93% across 1M+ daily events with 99.9% uptime, cutting manual review time from 3 minutes to 30 seconds. 🔹 Fine-tuned a multimodal Vision Language Model for recognizing 12 rare Indian traditional attires using PEFT and LoRA techniques. 🔹 Deployed a production-grade STT pipeline with punctuation restoration for English and Indic languages, achieving a 20% reduction in WER and up to 54% accuracy improvement at scale. 🔹 Deployed the Gemma 3 4B model using vLLM for news article summarization in the Jio's application, replacing third-party APIs to ensure consistent quality and lower cost. 🔹 Improved a custom object detection model's accuracy by 40%, enabling reliable identification of 101 distinct object classes. 🔹 Built automated data ingestion pipelines for daily vector database updates and implemented query caching, minimizing OpenAI API calls by 45%. Also mentor interns on AI/ML projects and lead technical discussions in daily scrum calls.

Assistant Manager - GET at Jio Platforms Limited (JPL), Dec 2023 - Jan 2024 · 2 mos, Hybrid, 🔹 Developed gender classification and age prediction models using Python to power user demographic profiling, enabling data-driven personalization across Jio's product ecosystem. 🔹 Designed and optimized end-to-end ML pipelines covering data preprocessing, feature engineering, model training, hyperparameter tuning, and evaluation for production-ready deployment. 🔹 Conducted comparative analysis of multiple ML algorithms (Logistic Regression, Random Forest, SVM, XGBoost) to identify optimal models for demographic prediction use cases.

Project Intern at Center of Excellence In High Performance Computing, Jan 2023 - Jul 2023 · 7 mos, Nagpur, Maharashtra, India

AIML Intern at SAMSU PRO CONSULTING SOLUTIONS PRIVATE LIMITED, Jan 2022 - Mar 2022 · 3 mos, Remote

Data Science and Business Analytics Intern at The Sparks Foundation, Aug 2021 - Sep 2021 · 2 mos, India

Education

Artificial Intelligence, Computer Science, G H Raisoni College of Engineering and Management(GHRCEMN) ,Nagpur

Public links

github: https://github.com/Fazle-hasan

medium: https://medium.com/@fazlehasan110