AI/ML Engineer · Deep Learning Researcher · Inference Engineer · Software Engineer — engineering intelligence at scale.
I design and ship intelligent systems where deep learning, large language models, and distributed systems engineering meet. My focus is the hard part of applied AI: making models fast, grounded, and reliable enough to run in production.
Lately I've gone deep on inference engineering — squeezing latency and cost out of LLM serving with quantization, KV-cache optimization, continuous batching, and speculative decoding. I build real-time voice agents with streaming ASR, barge-in, and sub-second turn latency, and I architect Graph RAG and vectorless RAG (paged document indexing) pipelines that retrieve by reasoning rather than nearest-neighbor lookups.
Underneath the AI work is a systems engineer. My background spans reinforcement learning, agentic orchestration (LangGraph, MCP), and cloud-native backend engineering in C++, Python, Java, and JavaScript — down to lock-free, per-core networking cores and SIMD-accelerated hot paths. I build adaptive systems that learn, reason, and scale.
Deep Learning · Large Language Models · Reinforcement Learning (PPO, SAC/DQN, RLHF, Reward Modeling) · Computer Vision (CNNs, ViT, CLIP, U-Net) · NLP · Generative AI · Recommender Systems · Agentic AI · Model Design, Fine-tuning, Evaluation & Interpretability
vLLM · TensorRT-LLM · Quantization (INT4/INT8) · Paged KV-Cache · Continuous (In-Flight) Batching · Speculative Decoding · CUDA Graphs · Triton · Sub-100ms TTFT serving
Streaming ASR (Whisper / faster-whisper) · Streaming TTS (Piper) · WebRTC / LiveKit · Silero VAD · Endpointing · Barge-in · Turn Detection · LLM Tool-Calling
Graph RAG (Knowledge Graphs, Neo4j / Cypher, multi-hop traversal, community summaries) · Vectorless / Paged RAG (hierarchical indexing, LLM routing, reranking) · Hybrid Retrieval · Vector Stores (OpenSearch, FAISS, Pinecone) · LangChain / LangGraph · MCP
REST / gRPC APIs · Microservices & Event-Driven Systems · FastAPI · Fastify / Node.js · Spring Boot · Next.js / React · PostgreSQL / MySQL · MongoDB / Redis · Kafka / RabbitMQ · Docker / Kubernetes · CI/CD · Observability · Caching & Rate Limiting
Distributed Systems · Multi-Threading · Lock-Free / Work-Stealing Schedulers · Parallel Processing · Shared-Nothing / Per-Core Architectures · MPI / OpenMP · CUDA / GPU · High-Performance Computing (300TB-scale HPC workloads)
C / C++23 · Systems Programming · io_uring / kqueue / IOCP · SIMD (AVX2 / NEON) · Memory Management · Zero-Copy I/O · High-Throughput Networking · ThreadSanitizer · Linux Internals · Profiling (perf, valgrind)
Languages: Python · C++ · C · Java · JavaScript / TypeScript · SQL ML/Infra: PyTorch · Hugging Face Transformers · CUDA · Ray · ONNX · Weights & Biases · scikit-learn / NumPy / pandas Cloud: AWS (EC2, S3, Lambda, SageMaker) · GCP (Vertex AI, BigQuery, Cloud Run) · Azure (OpenAI, Functions)
Purdue University, Indiana · 📅 January 2024 – May 2025 📈 GPA: 4.0 / 4.0 — Perfect GPA across 30 credits in just 3 semesters. 🏆 Top performer in AI/ML, High-Performance Computing, and Software Engineering coursework.
🧠 Coursework (all A-grade): Natural Language Processing · Computer Vision · Machine Learning · Operating Systems · High-Performance Computing · Full-Stack Development · Algorithm Design · Software Engineering · Human-Computer Interaction · Game Design & Simulation
National Institute of Technology, Warangal, India · 📅 August 2018 – May 2022 🏆 Ranked #9227 out of 1.2M+ candidates in JEE Mains (2018).
Always eager to collaborate on production AI, inference systems, and high-performance backends — or any project pushing toward new horizons. Reach out anytime!
Portfolio · LinkedIn · GitHub · Email
"Building the future, one inference at a time."