Vamshidhar Singh

Applied AI / Portfolio

Applied AI Engineer building production GenAI, speech AI and data systems in banking and fintech.

Open to AI/ML, Applied AI, Forward Deployed and Data roles · Open to relocate

View experience Download resume

500K+ calls/day92% intent resolution99.8% PII redaction accuracy$400K/yr saved

Experience

Bank of America

AI/ML Engineer · Jun 2024 to Present

Led architecture and end to end development of Erica, the bank's model agnostic AI virtual assistant integrating NLP, ML, GenAI/LLMs and banking services.
Built real time speech pipelines (gRPC, Kafka, Whisper/Conformer) ingesting 500K+ daily calls across 18K+ contact center reps at under 200ms.
Built multi agent workflows in LangGraph with state machine routing across intent classification, RAG retrieval and compliance review agents, reaching 92% intent resolution.
Built RAG pipelines with LangChain, Pinecone and ChromaDB over HR policy and compliance documents, making knowledge lookup 40% faster.
Enforced NeMo Guardrails, prompt injection defense, RBAC and real time PII redaction at 99.8% accuracy on financial data.

Erica: Bank of America's AI Virtual Assistant

Led architecture, 92% intent resolution

Real Time Speech AI and PII Redaction

500K+ calls/day, 99.8% PII redaction accuracy

Multi Agent Orchestration

LangGraph agents; related RAG pipelines enabled 40% faster knowledge lookup

Tata Consultancy Services

Machine Learning Engineer · Feb 2020 to Jul 2023

Leroy Merlin (Retail): GCP and Airflow pipelines processing 150M+ daily events at 5,000 events per second peak, 35% fewer pipeline failures, and segmentation models driving a 22% lift in click through rate.
Hub Group (Logistics): spatial routing and MILP optimization over 10M+ GPS points, 12% faster container turnaround, 25% better ETA accuracy and about $400K in yearly fuel savings.
Creatrix Campus (EdTech): transformer intent classification chatbot, LayoutLM document extraction with 50% faster admissions processing, and XGBoost retention models.
Banking and Telecom: IDFC FIRST Bank credit risk and fraud analytics over 45,000+ accounts, Deutsche Bank ETL pipelines, and Telefónica Germany network fault detection.

Hub Group Route Optimization

12% faster turnaround, about $400K/yr saved

Projects

Fintech Fraud Detection

XGBoost fraud scoring with FastAPI and a Streamlit dashboard.

XGBoost, FastAPI, Streamlit

View on GitHub ↗

Financial RAG Platform

RAG over 10K filings and earnings transcripts with FAISS and FastAPI.

RAG, FAISS, FastAPI

View on GitHub ↗

Financial Streaming Analytics

Kafka streaming with idempotent Postgres writes and a live dashboard.

Kafka, Postgres, Streamlit

View on GitHub ↗

Resume Job Matching System

Explainable resume to job scoring with spaCy and semantic search.

spaCy, FAISS, semantic search

View on GitHub ↗

AI Music Composer

LLM music generation with LangChain and Groq, deployed on GKE.

LangChain, Groq, GKE

View on GitHub ↗

AI Agents Travel Planner

Multi agent travel planning with CrewAI.

CrewAI, live search APIs

View on GitHub ↗

Education and Certifications

M.S. Business Analytics
Golden Gate University

B.Tech Information Technology
JNTU Hyderabad, India

AWS Certified Developer Associate (DVA C02)

Microsoft Certified: Azure AI Engineer Associate (AI 102)

Contact

The best way to reach me is email.

vamshidharsingh00@gmail.com

GitHub ↗LinkedIn ↗

Stack

Agents: LangGraph, LangChain, LlamaIndex, CrewAI, MCP, LiteLLM
Retrieval: RAG, Graph RAG, Pinecone, ChromaDB, Neo4j, Elasticsearch
Evaluation: RAGAS, DeepEval, LangSmith, Langfuse, MLflow, OpenTelemetry
ML: PyTorch, Fine tuning, RLHF, vLLM, Triton, XGBoost
Backend: Python, TypeScript, FastAPI, gRPC, SQL, pytest
Infra & Data: AWS, Kubernetes, Terraform, PySpark, Kafka, Databricks