ML Engineer and backend developer passionate about high-throughput data pipelines, LLM orchestration and AI Safety.

I am an Electronics and Communication Engineering student at NIT Patna (2022–2026) currently working as an ML Engineer at Bold Inc., where I have built an ultra-lightweight infrastructure working as an AI Gateway and data transport systems.

My top works:

• 10K concurrent requests handled at sub-1ms p99 latency via a custom lock-free Go AI gateway.
• 6.3GB/s transfer speeds achieved on a high data processing pipeline in Go using consumer hardware.
• Selected for FAR.ai global AI Safety Summit in London, UK (2026).
• Top 10% selection for the global M2L School ML summit in Split, Croatia (2025).
• Paper on hybrid KG-RAG reasoning during my research internship at IIIT Hyderabad.
• Winner of the IIT ISM AI Challenge (built an optimized OCR pipeline).

• Entropy of Thinking Chain-of-Thought as Contextual Stabilization and Associative Retrieval: Why may LLMs not be the intelligent engines which we think they are? This is why we can’t ask them to be safe and expect them to be safe.

• Efficient LLMs via Switchable and Dynamic Quantization Quantization of LLMs and testing the behaviour if an LLM is finetuned to be safe at a higher precision and is quantized.

• Distributed Pipeline Parallelism — High-performance distributed orchestration strategies.
• High Data Processing Pipeline in Go — Eliminating thread contention with O(1) worker pools.
• Developer Token Observability Platform — Go/Python telemetry system with sub-18ms P99 latency.
• Automatic Document Ingestion Engine — LLM-powered invoice screening processing 1K+ documents/week.