Subhash Chandra Pal

With over 15 years of experience in software architecture and intelligent systems, I bring 8+ years of hands-on expertise in Deep Learning and Machine Learning, with a specialized focus on Generative AI, Reinforcement Learning, and Model Quantization. Over the past 5+ years, I’ve worked extensively on designing small and large-scale language models (SML/LLM), applying them to domains like conversational AI, computer vision, and edge intelligence.

🤖 Generative AI & Language Models

I specialize in building and optimizing LLMs (BERT, GPT, Transformer architectures) with a strong foundation in quantization, fine-tuning, and deployment. I’ve developed solutions for sentiment analysis, multi-lingual transformers, image generation, and agent-based systems powered by Agentic AI and RAG (Retrieval-Augmented Generation).

🧠 Reinforcement Learning & Adaptive Systems

Certified in Reinforcement Learning (IITM), I design and implement adaptive learning environments for autonomous decision-making, real-time tracking, and intelligent control systems, particularly in embedded and robotic environments.

⚙️ Programming & System Optimization

I’m highly proficient in Rust, Python,Golang ,C++ and Julia with an emphasis on performance engineering. My work in Rust and C++ is focused on high-throughput, low-latency AI workloads, model deployment on edge, and AI kernel development for HPC and embedded platforms.

🚀 Embedded Systems & Edge AI

With a strong foundation in Embedded Linux, QNX, and Yocto, I develop real-time AI systems for drones, industrial cameras, and autonomous vehicles. My work integrates sensor fusion, gesture recognition, and computer vision into optimized edge deployments.

🌐 Cloud AI & Distributed Systems

Experienced in AWS, Azure, and GCP (Vertex AI), I design and deploy scalable ML workflows and microservices using gRPC, Docker, and Kubernetes. I’ve successfully built real-time ML pipelines, from training at scale to deploying quantized models for efficient inference.