Anudeep Sri Bathina

I build production-grade GenAI systems that survive real traffic, real constraints, and real scrutiny - with explicit retrieval boundaries, rigorous evaluation, and measurable reliability.

Build-ready

Strict lint, type-check, static export

Architecture-first

RAG, evals, safety, deployment trade-offs

Proof surfaces

Projects, notes, publications, resume

Agentic AILangGraph · CrewAI · MCP
RAG SystemsAzure AI Search · ChromaDB · GraphRAG
MLOpsFastAPI · Docker · LangSmith
Computer VisionYOLOv8 · ResNet · 94.1% mAP
AT&TCapgeminiCognizant

Current

AI Architect

AT&T

11+

Years in prod

1K+

Mentored

Published - CRC Press
Speaker - BSBI Berlin
KaggleX Fellow - Google

15K+

Enterprise GenAI users

2.4s

p95 RAG @ $0.42/1K

For Recruiters

Open to Staff+ AI Engineer / AI Platform Engineer roles (L5–L6)

Particularly interested in teams building reliable, observable, and cost-efficient GenAI platforms at scale - with strong emphasis on RAG architectures, evaluation systems, agentic workflows, and production guardrails.

Currently exploring

Reviewed May 2026

Agentic AI orchestration with LangGraph + MCP
Multi-modal RAG with vision models
AI safety evaluation frameworks
Building ClinIQ v2 with department-scoped retrieval

Not another AI demo.

I bring deep research rigor and battle-hardened production engineering. Models that work in notebooks don't interest me. I build systems that hold up under real traffic, real data, and real deadlines.

0+

years shipping production AI

0

engineers led across projects

AT&T/AI Architect
Capgemini/Tech Lead
Cognizant/Engineer
Full career timeline
2024

AI Architect

AT&T
2021

M.S. Computer Science

UMass Dartmouth
2019

Technical Lead

Capgemini
2015

Data Science Engineer

Cognizant
2011

B.Tech CS

VIT University

Key Impact at a Glance

Production results from enterprise GenAI and ML systems

15K+
Users on production GenAI platform
AT&T
4h → <90s
Policy query resolution time
Reduced via RAG + automation
$2.3M
Retention savings delivered
Through ML churn models
8 engineers
Team led
Increased deployment frequency 4×
2.4s p95
RAG latency @ $0.42/1K requests
Production Azure deployment
97.2%
Document extraction accuracy
50K+ pages/month with PII safety

Production Toolkit

The tools, patterns, and architectures I use to ship reliable AI systems at enterprise scale.

Agentic AI & Orchestration

LangGraph/CrewAIMCP (Anthropic)OpenAI o3/o4-miniClaude 4/Gemini 2.5

RAG & Knowledge Systems

Advanced RAG PipelinesChromaDB/Pinecone/QdrantGraphRAGCohere/BGE/ColBERT

AI Safety & Eval

LangSmith/LangfuseGuardrails & GradingPresidio (PII/PHI)Red-teaming & Evals

AI-Native Dev

Claude Code/Codex CLICursor/WindsurfGitHub CopilotA2A Protocol

Cloud & Infra

Python & FastAPIAzure AI/AWS BedrockDocker & KubernetesCI/CD & MLOps

Data Engineering

PySpark/DatabricksSQL & dbtAirflow/PrefectKafka & Streaming

Open Source Reference Implementations

Production patterns and architectures I've released for teams building dependable AI systems at scale.

ClinIQ

FEATUREDOPEN SOURCE
Production Healthcare RAG with department-scoped retrieval & PII boundaries
Healthcare RAG · citations · PII safety

Enterprise Healthcare RAG reference implementation with scoped retrieval, clinical citations, and production-grade privacy controls.

LangGraphLangChainFastAPIOpenAIChromaDBPresidio

EvidenceIQ

OPEN SOURCE
Evidence-grounded RAG with source ranking & review workflows
evidence retrieval · source ranking · citations

Evidence-grounded RAG pattern for retrieving, ranking, and explaining source-backed findings with strong citation guarantees.

RAGEvaluationCitationsPythonLLM

InferIQ

OPEN SOURCE
LLM output quality gates & regression checks for production release
eval gates · output checks · regression detection

Production-grade LLM evaluation harness for checking output quality, safety, and regression before models reach users or CI/CD.

LLM EvaluationQuality GatesPythonObservability

Code Migration Assistant

OPEN SOURCE
Enterprise code migration with AI risk analysis and dependency visualization
code migration · risk assessment · compliance

Enterprise-grade, security-first code migration tool with AI-powered risk assessment, visual dependency planning, and compliance scanning.

PythonASTNetworkXRAGOpenAISecurity

Annapurna-AI

OPEN SOURCE
Domain-specific agentic applications with model routing & cultural context
multimodal · model routing · cultural AI

Culture-aware, India-first AI meal planner and grocery assistant for South Indian vegetarian cooking with evidence-grounded recommendations.

Next.jsFastAPIGeminiLiteLLMSQLModel

Teaching & impact

Mentoring AI practitioners across 20+ countries through hands-on sessions, guest lectures, and workshops.

0+

Years building AI

0+

Learners reached

0M+

Business value delivered

0+

Countries reached

Anudeep is incredibly insightful, listening carefully and offering technical yet straightforward comments that are truly beneficial.

Michael

Freelance Developer

via ADPList
UMass DartmouthNKUVIT UniversityDeepLearning.AIBSBI BerlinGreat LearningWileyVerzeoAnalytics VidhyaKaggle

Let's talk.

AI consulting, speaking, mentorship, or research collaboration.