M.TECH FOCUS · BITS PILANI
Intelligence needs
infrastructure.
I’m building depth across modern AI while applying a production engineer’s lens: data quality, reproducibility, evaluation, observability, cost, security, and dependable deployment.
01 / GENERATIVE SYSTEMSGenAI
LLM application patterns, prompt and context engineering, embeddings, vector search, agentic workflows, evaluation, guardrails, and production inference.
LLMsAgentsEvaluationInference
02 / MODEL SERVINGInference Engineering
Production model serving optimized for predictable latency, high throughput, efficient compute, and observable behavior under real workloads.
vLLMBatchingQuantizationKV cacheGPU utilizationLatency
03 / RETRIEVALRAG + CRAG
Retrieval pipelines, corrective retrieval, reranking, grounding, citations, chunking strategies, and reliable knowledge systems.
04 / LEARNINGDeep Learning
Neural networks, optimization, representation learning, transformers, CNNs, training dynamics, and model evaluation.
05 / DECISION MAKINGReinforcement Learning
MDPs, value and policy methods, exploration, reward design, Q-learning, policy gradients, and sequential decision systems.
06 / PERCEPTIONComputer Vision
Image classification, detection, segmentation, visual representation learning, multimodal systems, and efficient vision inference.
07 / EMBODIED AIRobotics
Perception-to-action loops, planning, control, localization, sensor fusion, simulation, and the software infrastructure behind autonomous systems.
08 / NEXT COMPUTEQuantum Computing
Quantum information foundations, qubits, gates, circuits, measurement, core algorithms, and hybrid quantum-classical computation.
09 / FOUNDATIONSMathematics
The layer beneath every model and system.
Linear algebraProbabilityStatisticsCalculusOptimizationNumerical methods