AI / ML ENGINEER

Likith
Chilakalapadi Prabhu

Building reliable AI systems from retrieval to inference.

I design and build production-grade Generative AI, RAG, agentic workflows, LLM evaluation platforms, and scalable inference infrastructure.

Python · PyTorch · RAG · LangGraph · FastAPI · AWS · Kubernetes
View Projects ResumeLinkedIn
AGENT
RAG
LLM
EVAL
PRODUCTIONAI SYSTEMRELIABLE · SCALABLE
5+Years in AI / ML
34%Inference latency reduction
35K+Enterprise users supported
20+Evaluation configurations benchmarked

ABOUT

Engineering AI systems that work beyond the demo.

I focus on the full production AI lifecycle: retrieval, orchestration, model serving, evaluation, observability, and infrastructure.

My work centers on building AI systems that are not only accurate, but also measurable, scalable, reliable, and maintainable in real-world environments.

EXPERIENCE

Production AI from models to infrastructure.

NVIDIAMar 2025 — Present

AI/ML Engineer

Building production AI systems across multi-agent orchestration, enterprise RAG, distributed GPU inference, LLMOps, and cloud-native infrastructure.

  • Supported enterprise AI experiences serving 35K+ employees, improving response accuracy by 19%.
  • Improved RAG retrieval quality by 21% through retrieval, reranking, and evaluation improvements.
  • Reduced inference latency by 34% and infrastructure cost by 28% through GPU-serving and deployment optimization.
AccentureOct 2020 — Jul 2024

Machine Learning Engineer

Built production AI/ML platforms spanning RAG, scalable Python services, MLOps, event-driven pipelines, cloud infrastructure, and enterprise integrations.

  • Developed AI/ML systems supporting products with up to 420K users.
  • Improved RAG retrieval performance by 21% across enterprise knowledge workflows.
  • Reduced inference latency by 33% and infrastructure cost by 29% through serving and cloud optimization.

SELECTED WORK

Evaluation infrastructure for production LLM systems.

A production-oriented evaluation platform for measuring LLM quality, catching regressions before release, and comparing model, prompt, and retrieval configurations at scale.

FLAGSHIP PROJECT

EvalForge

LLM Evaluation & Regression Testing Platform

01

Automated evaluation platform for benchmarking model, prompt, and retrieval configurations across correctness, groundedness, citation accuracy, hallucination rate, latency, and cost.

01Test Dataset

Versioned evaluation cases and expected behavior.

→
02LLM Pipeline

Model, prompt, and retrieval configurations.

→
03Evaluation

Deterministic and LLM-as-a-judge scoring.

→
04Regression Gate

Blocks releases when quality thresholds fail.

93%intentional regressions detected
72%less manual evaluation effort
<250msP95 API latency
50+concurrent evaluation jobs
PythonFastAPIPyTorchMLflowPostgreSQLRedisAWS
LLM Evaluation Platform

TECHNICAL STACK

Technical depth across the production AI lifecycle.

Generative AI & Agents

LLMs · Agentic AI · RAG · LangGraph · MCP · Tool Calling

ML & NLP

PyTorch · Hugging Face · NVIDIA NeMo · Embeddings · Reranking

Inference & Serving

TensorRT-LLM · Triton · vLLM · CUDA · FlashAttention

Backend & Data

FastAPI · gRPC · Kafka · Redis · PostgreSQL · Spark

Cloud & MLOps

AWS · Kubernetes · Docker · Terraform · MLflow · GitHub Actions

Evaluation & Reliability

LLM Evaluation · DeepEval · Observability · Guardrails · Tracing

EDUCATION

Master of Science in Computer Science

University at Buffalo, New York

Bachelor of Technology in Software Engineering

Vellore Institute of Technology

CERTIFICATIONS

  • NVIDIA Certified Associate — Generative AI LLMs
  • AWS Certified Machine Learning Engineer — Associate
  • Databricks Certified Generative AI Engineer Associate
  • Certified Kubernetes Administrator (CKA)

CONTACT

Interested in building production AI systems?

Open to AI/ML engineering opportunities and technically ambitious teams.

Email meLinkedIn