Manojkumar Padmanaban

About Manojkumar Padmanaban

I’m a Senior Consultant – AI & Data at Deloitte India, with 7+ years of total experience driving data science and AI solutions across Automotive, FMCG, Pharma, and Transportation industries. I’ve led initiatives that reduced operational costs by up to 40% and accelerated QA processes by 50% through AI-powered automation. I design and deploy end-to-end Generative AI solutions, including supervised fine-tuning (SFT), parameter-efficient fine-tuning (PEFT), RLHF, and AI agents. Expert in multi-GPU training, model serving, and scaling systems for enterprise applications like intelligent document processing (processing documents daily with 95% accuracy) and automated test-case generation (generating 500+ cases per sprint, cutting manual effort by 60%). Specializing in Generative AI, Computer Vision, Deep Learning, and ML Ops, I deploy scalable solutions using Python, TensorFlow, PyTorch, and cloud platforms like GCP (Vertex AI) and Azure. Recent projects include leading enterprise-scale GenAI initiatives, such as fine-tuned model training and deployment with LoRA adapters, enabling faster compliance, audit, and QA automation. With a blend of engineering expertise and business acumen, I’ve empowered enterprises to achieve digital transformation—e.g, optimizing supply chains in FMCG to boost efficiency by 30% via predictive analytics.Outside work, I’m passionate about open-source contributions, including a GitHub repo on PEFT techniques that’s garnered 500+ stars, and mentoring aspiring AI professionals through workshops. Academically, I hold: • PGP in Data Science from Bangalore (2024) • B.Tech in Information Technology (2018)I’m always open to collaborating on AI/ML projects, product ideation, and innovations in Generative AI and Applied ML. Let’s connect!

Experience

1. Senior Consultant - Generative Ai

   Deloitte

   May 2024 — Present · Bengaluru, IN

   Developing end-to-end Generative AI solutions utilizing custom fine-tuned Large Language Models (LLMs) on private datasets, ensuring data security through on-premise GPU clusters.• Automating test-case and test script generation by implementing Automated Test Case Generation techniques, leveraging fine-tuned Text and Code Generative LLMs.• Adapting open-source foundational LLMs such as Falcon, LLaMA2, and Mistral to domain-specific data through Domain Adaptation strategies, addressing use cases including Legal e-Discovery, Internal Audit, and Regulatory Compliance.• Exploring state-of-the-art Generative Models (LLM) like LLaMA 3.1, Qwen and Mistral to solve diverse client challenges, ensuring solutions are tailored to specific industry needs.• Implementing parameter-efficient fine-tuning methods such as Low-Rank Adaptation (LoRA) and Quantized LoRA (qLoRA) to optimize model performance and resource utilization during instruction tuning.

Education

• Sona College of Technology

  Bachelor's degree

  2014 — 2018

• International Institute of Information Technology Bangalore

  Postgraduate Degree

  2023 — 2024