AI Systems Engineer

LLM inference and serving runtimes · robotics perception · high-throughput backend infrastructure

I build systems that make AI products usable under real constraints: latency, throughput, observability, deployment reliability, and evaluation quality. My strongest work sits at the intersection of LLM infrastructure, cloud-native backend services, and robotics/autonomy pipelines.

HelixServe LLM serving runtime on GCP NVIDIA L4 with paged KV cache, continuous batching, prefix caching, CUDA Graph decode, and benchmark instrumentation.

ManoVarta Controller-led multilingual mental-health GenAI system for English, Hindi, and Hinglish PHQ-9/GAD-7 item-level assessment with evidence extraction and safety routing.

SRE-Nidaan Incident response copilot using Next.js, FastAPI, vLLM, telemetry grounding, runbook retrieval, remediation gating, and analyst feedback loops.

Nervaflow Intelligence Cloud-native decision engine for supply operations using Vertex AI Search, conversational APIs, BigQuery pipelines, operational traces, and cost attribution.

Distributed Robotics and Networked Embedded Sensing Lab — Research Aide, Robotics Systems Engineering

• Reduced mapping drift to under 1.2% across 500m of GPS-denied environments by building a ROS2 visual SLAM pipeline for Boston Dynamics Spot with Gaussian Splatting integration.
• Improved LiDAR and VIO fusion for low-texture subterranean navigation, reducing trajectory estimation error by 30% while sustaining 20 Hz real-time state estimation.

Tata Elxsi — Software Engineer Intern

• Migrated autonomous vehicle software from ROS1 to ROS2 and tuned DDS QoS behavior in CARLA, reducing inter-module latency by 40% and achieving sub-100 ms communication latency.
• Built an Extended Kalman Filter-based vehicle state estimator and safety-constrained route planner with deterministic state transitions, bounded-latency path generation, and validation hooks.

LLMate.ai — Backend Engineer Intern

• Built asynchronous backend services with Spring Boot and RabbitMQ for production data workflows, reducing p95 response latency by 40%.
• Deployed a GPT-3.5-based text-to-SQL workflow over 50,000 structured records and set up Docker/GitHub Actions CI/CD.

• Languages: Python, C++, Java, SQL
• AI systems: LLM inference, model serving, vLLM, Vertex AI, QLoRA, RLHF, CUDA Graphs, prompt evaluation
• Backend and cloud: FastAPI, Spring Boot, Docker, GCP, Cloud Run, BigQuery, RabbitMQ, distributed systems, observability, CI/CD
• Robotics and autonomy: ROS2, SLAM, LiDAR/VIO sensor fusion, CARLA, state estimation

I am looking for AI infrastructure, backend/product engineering, ML systems, or robotics/autonomy roles where I can own systems end to end: from low-level performance and reliability work to shipped user-facing demos.