AI/ML ENGINEER · 3+ YRS · LOS ANGELES

I BUILD AI THAT ACTUALLY SHIPS —
FROM FIRST PROMPT TO PROD.

Hi, I'm Chandar Rathi. I design, build, and run GenAI and ML systems end-to-end — RAG pipelines, LLM apps, and forecasting models on AWS and Azure — for fintech and healthcare, where "it mostly works" doesn't cut it.

See the projects ↓ Download résumé
+28%
RETRIEVAL RELEVANCE via hybrid retrieval + reranking + embedding optimization
+22%
FORECAST ACCURACY with Prophet + LightGBM demand models
+25%
API RESPONSE SPEED via Redis caching over PostgreSQL
99.9%
SERVICE AVAILABILITY with CI/CD + CloudWatch monitoring

Where I've shipped

EXPERIENCE

PayPal

MAR 2025 — PRESENT · CA, USA

AI/ML Engineer

ENTERPRISE AI PAYMENT SUPPORT ASSISTANT

Deloitte (end-client: CVS Health)

FEB 2021 — DEC 2023 · INDIA

ML Engineer

ENTERPRISE PHARMACY DEMAND FORECASTING PLATFORM

The desk

PROJECTS
↯ these cards are physically draggable — go on, make a mess

RAG Ops — Production RAG Assistant

LANGCHAINPGVECTORFASTAPIREACT

Open-sourced, from-scratch RAG assistant built to benchmark retrieval strategies — a reproducible naive vs. hybrid-plus-reranking comparison on a 500-question eval set, with a streaming Next.js UI and a Ragas dashboard exposing faithfulness and recall per query.

Serves quantized Llama 3 / Mistral on a single GPU with vLLM — near-zero per-query cost, P95 under 800 ms.

NDCG@5: 0.64 → 0.89 View repo ↗

LLM Observability & Cost Platform

OTELCLICKHOUSEGRAFANAFASTAPI

Self-hostable tracing for AI apps — every LLM request end to end: prompt, retrieved context, tool calls, tokens, latency, cost. Traces land in ClickHouse; per-route cost and latency surface in Grafana with automated alerts when spend or quality drifts.

Ships as drop-in middleware that instruments any LangChain or LangGraph app in a few lines, with a public demo and docs.

Debug + budget AI, not guess View repo ↗

The toolbox

SKILLS

GENERATIVE AI

LLMsRAGLangChainprompt engineeringhybrid retrievalsemantic searchMCPtool callingRagasLangSmithpgvector

MACHINE LEARNING

Scikit-learnLightGBMProphetfeature engineeringpredictive modelingtime-series forecastingPySpark

CLOUD & MLOPS

AWS BedrockAmazon EKSS3CloudWatchAzure MLDatabricksAzure Data FactoryADLS Gen2DockerMLflowGitHub ActionsCI/CD

BACKEND & DATA

PythonSQLFastAPIREST APIsmicroservicesPostgreSQLRedisAzure SQLGitAgile/Scrum

Take the paperwork

DOWNLOADS

Résumé

Everything above, one tidy PDF. ATS-friendly, recruiter-approved, zero rotated stickers.

Download résumé ↓

Cover letter

The why behind the what — who I am, how I work, and what I'm looking for next.

Download cover letter ↓
Prefer email? Both land in your inbox — just ask below ↓