Dhirendra Choudhary — AI Researcher

Building a demo is easy. Making AI reliable in production is hard. I am a Senior AI Engineer focused on that specific gap: taking Generative AI from "experimental" to "enterprise-grade." My work focuses on the engineering challenges of Autonomous Systems—specifically Latency, Accuracy, and Cost. I don't just integrate APIs; I design the cognitive architectures that allow models to reason correctly (Agentic AI) and fail gracefully. What I build (The Stack): - Agentic Orchestration: Currently building complex, stateful workflows using LangGraph and CrewAI to handle multi-step reasoning (Time-series planning, Analytics). - Advanced Retrieval: Moving beyond naive vectors. I have designed GraphRAG systems (using Neo4j) to solve complex relationship mapping in procurement and analytics. - Model Optimization: I specialize in making models cheaper and faster. This involves Fine-tuning SLMs (Llama-3, Mistral) using QLoRA/PEFT and optimizing inference via KV-cache and quantization. My Background: I bridge the gap between academic research and industrial application. With a Master’s in Computer Science and published research in pattern recognition (Sign Language/EEG), I approach AI with a rigorous, scientific mindset. Currently, I’m helping build the next generation of intelligent systems at Mphasis. What I write about: I share my daily learnings on the "Modern AI Stack"—moving beyond basic implementations to architecting robust, scalable systems. If you are interested in the engineering side of AI (Evals, Quantization, Agents), let’s connect.

Stackforce AI infers this person is a Senior AI Engineer specializing in enterprise-grade AI solutions and model optimization.

Location: Bengaluru, Karnataka, India

Experience: 8 yrs