Open to opportunities — AI/ML Engineer at Truist Bank

Raviteja Annam

|

I build production agentic AI systems that solve real problems. From multi-agent RAG pipelines processing 10K+ documents to fraud detection at 50M+ transactions — I ship AI that works at scale.

7+
Years Experience
6
Deployed Projects
$6.9M
Business Impact
99.7%
Uptime SLA
scroll
01

About Me

I'm an AI/ML Engineer with 7+ years building production systems that ship. My path — cloud engineering to fintech data science to agentic AI — gave me something rare: I understand infrastructure, models, and the humans who use them. Every project I build is live, deployed, and solves a real problem.

Agentic AI Systems

Multi-agent orchestration with LangGraph, HITL safety layers, and production-grade RAG pipelines at Truist Bank.

ML at Scale

Fraud detection on 50M+ monthly transactions, real-time ML platform processing 0.5B daily events at Razorpay.

AI for Impact

Built GlioblastomaGPT after a personal family diagnosis — medical AI that translates research into hope.

Production Impact

0%
Lookup Reduction
RAG at Truist
0%
Error Reduction
HITL Review Layer
$0M
Chargebacks Saved
Fraud Detection
$0M
GMV Recovered
Payment Routing
02

Experience

From cloud infrastructure to fintech ML to agentic AI — a trajectory of increasing complexity and impact.

Truist Bank

AI/ML Engineer — Agentic AI & LLM Infrastructure

Jan 2025 — Present
Charlotte, NC (Remote)
~55%
Lookup Reduction
~40%
Fewer AI Errors
6+
Teams Adopted MCP
99.7%
Uptime SLA
  • Built RAG pipeline over 10,000+ compliance docs with LangChain + Pinecone, saving ~30 hrs/week
  • Designed HITL review layer in LangGraph reducing erroneous AI decisions by ~40%
  • Led MCP tool integrations — per-tool onboarding from 2 days to under 4 hours
  • Built Ragas + LangSmith eval pipeline with 500+ golden queries — 12-pt faithfulness improvement
  • Containerized LLM services on AWS EKS with Terraform, deployment time 2 days to under 3 hours
LangGraphLangChainGPT-4oAWS BedrockPineconeFastMCPRagasLangSmithDockerEKSTerraform

Razorpay

Senior Data Scientist — ML Platform & Risk Engineering

Feb 2019 — Jul 2023
Bangalore, India
$2.1M
Chargebacks Saved
$4.8M
GMV Recovered
0.5B
Daily Events
94%
Fraud Precision
  • Built XGBoost fraud detection on 50M+ monthly transactions — 94% precision, 91% recall
  • Added GNN layer for fraud ring detection, +28% improvement
  • Thompson Sampling MAB for payment routing — 3.2% lift, ~$4.8M annual GMV recovered
  • NLP ticket classifier across 15K+ monthly tickets, 60% routing time reduction
  • Architected real-time ML platform (Mitra) — Flink + Kafka, sub-200ms latency
  • Standardized MLOps with MLflow across 8-person team, 50% faster model re-validation
XGBoostPyTorch GeometricGNNThompson SamplingSageMakerSHAPLIMEspaCyKafkaFlinkDelta LakeMLflow

MUV Technologies

Cloud Support Engineer

Jun 2017 — Jan 2019
Hyderabad, India
50+
EC2 Instances
~15h
Weekly Time Saved
  • Managed 50+ AWS EC2 instances with Python and Bash automation
  • Led zero-downtime ERP and CRM migrations on AWS infrastructure
AWS EC2PythonBashCloud Infrastructure

Oklahoma City University

M.S. Computer Science — AI/ML Specialization

Aug 2023 — May 2025
GPA: 3.9 / 4.0
Generative AILLMsDeep LearningReinforcement LearningNLPDistributed Systems
03

Projects

6 deployed projects with live demos. Every one solves a real problem — from medical AI built out of personal need to fintech ML at production scale.

Medical AI

GlioblastomaGPT

Multi-Agent Medical AI Research Assistant

Built after my father's brain cancer diagnosis. A 5-agent LangGraph pipeline that translates complex oncology research into language patients and caregivers can actually understand. Hybrid RAG over 50,000+ PubMed abstracts with hard safety layers.

0.84
Faithfulness
0.89
Relevancy
50K+
PubMed Docs
5
LangGraph Agents
LangGraphLangChainGPT-4oscispaCyBGE-M3PineconeBM25RagasFastMCPDocker
Agentic AI

AgentPulse

Autonomous AI Newsletter Pipeline

Fully autonomous personalized newsletter — 6 FastMCP servers discover content from GitHub, arXiv, Hacker News, YouTube, and Substack. SM-2 spaced repetition learns your preferences over time.

6
MCP Servers
6
LangGraph Agents
5
Data Sources
FastMCPLangGraphLangChainGPT-4oChromaDBSQLiteSendGridDocker
FinTech

FinSight

Agentic SEC Filing Analysis

RAG over SEC 10-K and 10-Q filings with financial NER and agentic query decomposition. Mirrors the financial AI workflows built at Truist Bank.

RAG
Architecture
NER
Entity Extraction
LangChainFAISSGPT-4ospaCyFastAPIPython
FinTech ML

Fraud Pattern Detector

XGBoost + GNN Ensemble with Explainability

Real-time fraud detection system combining XGBoost for tabular features and Graph Neural Networks for fraud ring detection. Full SHAP explainability dashboard shows exactly why each prediction was made.

94%
Precision
91%
Recall
$2.1M
Impact
XGBoostPyTorch GeometricGNNSHAPScikit-learnStreamlit
FinTech ML

Credit Risk Engine

BNPL/Merchant Scoring with LIME + SHAP

End-to-end credit scoring pipeline for BNPL and merchant lending. Dual explainability with LIME (local) and SHAP (global) for regulatory compliance.

+18%
Approval Lift
LIME
Local Explain
SHAP
Global Explain
XGBoostLIMESHAPScikit-learnSageMakerStreamlit
NLP

CineEnglish

NLP Movie-Based Language Learning

Uses NLP to score films by linguistic difficulty and builds a personalized curriculum matched to a learner's level. Semantic similarity for film recommendations.

NLP
Difficulty Scoring
ML
Recommendations
Sentence TransformersScikit-learnNLPStreamlitPython
04

Skills

Every skill listed here has been used in production — not just tutorials or certifications.

Agentic AI & LLMs

LangGraphExpert
LangChainExpert
RAGExpert
MCP / FastMCPExpert
OpenAI GPT-4oExpert
HITL SystemsExpert
Ragas + LangSmithExpert
AWS BedrockIntermediate
GroqIntermediate
PineconeExpert
FAISSAdvanced
ChromaDBAdvanced
🧠

ML & Deep Learning

XGBoostExpert
PyTorchAdvanced
GNN (PyTorch Geometric)Advanced
Scikit-learnExpert
SHAP / LIMEExpert
Thompson Sampling MABAdvanced
MLflowAdvanced
💬

NLP & Embeddings

Sentence TransformersAdvanced
BGE-M3Advanced
scispaCy / spaCyAdvanced
HuggingFaceAdvanced
BM25 (Hybrid Retrieval)Advanced

Cloud & MLOps

AWS (SageMaker, EKS, Lambda, S3, Bedrock)Advanced
DockerExpert
KubernetesAdvanced
TerraformIntermediate
FastAPIExpert
GitHub ActionsIntermediate
📊

Data Engineering

Apache Kafka + FlinkAdvanced
Delta LakeAdvanced
SQL / Pandas / NumPyExpert
Supabase + pgvectorAdvanced
💻

Languages & Frameworks

Python (7+ years)Expert
TypeScript / Next.jsIntermediate
SQLExpert

Best Fit Roles

Agentic AI EngineerLLM EngineerSenior ML EngineerAI Platform EngineerNLP EngineerGenerative AI EngineerApplied Scientist
05

Get in Touch

Let's build something great together.

I'm looking for my next role in agentic AI, LLM infrastructure, or applied ML. If you're building something that matters — in healthcare, fintech, or any domain where AI creates genuine value — I'd love to talk.

Schedule a Call

What I'm Looking For

  • Multi-agent systems or RAG pipelines in production
  • Teams that take evals and reliability seriously
  • Problems in healthcare, fintech, or infrastructure
  • Remote-friendly, Bay Area, or open to US relocation
Email
ravitejas1596@gmail.com
Phone
+1 (405) 402-9977
LinkedIn
linkedin.com/in/ravitejas1596
GitHub
github.com/Ravitejas1596
Calendly
Book a 30-min Call
Currently Available

24-48 hrs for intro calls, 2-3 days for technical interviews. F-1 OPT + STEM extension through ~mid-2028.