Building Production AI Systems That Deliver Real Impact
AI/ML Engineer with 8+ years of experience building and deploying machine learning and generative AI systems in production on AWS. Past two years focused on GenAI - RAG pipelines, multi-agent workflows with LangGraph, and LLM integration using AWS Bedrock and OpenAI. Built systems processing 50K+ documents, improved forecasting accuracy by 30-45%, and deployed MCP-based connectors for secure agent-to-data access. Hands-on across the full AI lifecycle - experiment tracking (MLflow), containerized deployment (Docker, ECS), and production monitoring (CloudWatch). AWS certified with deep experience across SageMaker, Bedrock, ECS, Lambda, API Gateway, and IAM.
Enterprise RAG platform ingesting and retrieving from large document corpus
10+
Teams Enabled
Worked cross-functionally to operationalize AI across product, architecture, and business teams
8+ yrs
AI/ML Experience
Building and deploying production ML and GenAI systems on AWS
Experience
AI/ML Engineer | Generative AI & Agentic Systems
AUConnects LLC, Atlanta, GA | August 2025 - Present
Architected and deployed an enterprise RAG platform processing 50K+ documents using LangChain, Pinecone, and FAISS - reduced retrieval time by ~60% for operations and finance teams
Designed agentic AI workflows using LangGraph with multi-step reasoning, task delegation, tool invocation, and response validation
Deployed LLM-powered agent workflows using AWS Bedrock (Claude) and OpenAI GPT-4 with tool-calling, structured outputs, and safety guardrails
Built RAG pipelines using OpenAI and Titan embeddings with chunking strategies, re-ranking, and metadata enrichment
Integrated Neo4j knowledge graphs for hybrid KG + vector retrieval (KG-RAG) to improve factual consistency
Designed serverless GenAI APIs using AWS Lambda and API Gateway for low-latency internal tooling
Implemented blue-green deployment patterns for containerized AI services on ECS with autoscaling policies
Designed MCP-based connectors and tool schemas enabling agents to securely invoke APIs without exposing credentials
Containerized services with Docker and set up CI/CD pipelines using GitHub Actions and Amazon ECR
This enterprise pipeline powers self-service AI at AUConnects, combining retrieval, orchestration, and agentic automation to reduce information search time and support faster decision-making across the organization.
Carrier HVAC, Charlotte, NC | January 2024 - August 2025
LLM-Powered Document Analysis
Led development using AWS Bedrock (Claude) and LangChain - operations teams could query technical manuals and SOPs in natural language, reducing manual review by ~40%
RAG Pipeline for HVAC Docs
Built RAG pipeline over internal documentation using Titan embeddings, FAISS, and LangChain with recursive chunking and metadata filtering
Prompt Engineering & Extraction
Developed prompt engineering workflows for structured extraction from supplier reports and warranty claims
PowerBI AI Integration
Integrated LLM-based summarization into PowerBI dashboards via FastAPI endpoints for AI-generated insights
Ensemble Forecasting Models
Built and deployed XGBoost, LightGBM, Prophet models improving demand planning accuracy by ~30-45%
Anomaly Detection & MLOps
Implemented anomaly detection pipelines; operationalized ML models using Databricks, MLflow, and Docker with version-controlled deployment workflows
Sabnext Solutions, Wichita, KS | June 2023 – December 2023
Built gradient boosting models (XGBoost, CatBoost) for churn prediction, lead scoring, and revenue forecasting - used by sales leadership for quarterly planning
Developed customer segmentation models using RFM analysis and K-Means clustering to identify high-value cohorts
Implemented anomaly detection pipelines for CRM data quality monitoring, catching data drift and ingestion issues early
Deployed FastAPI-based scoring services integrated with Salesforce for real-time predictions on incoming leads
Packaged inference services with Docker and deployed on AWS EC2 with S3 artifact storage and CloudWatch logging
InterviewAI - Full-Stack AI Mock Interview Platform
FastAPI + Streamlit | Production-Grade
Built a production-grade AI mock interview platform with a decoupled architecture using FastAPI backend and Streamlit frontend, deployed on Streamlit Cloud
Engineered a conversational interview engine powered by OpenAI GPT-4 with dynamic question generation, multi-turn follow-ups, and structured performance scoring
Designed RESTful API endpoints for session management, interview orchestration, and real-time feedback delivery with stateful conversation tracking
Containerized the backend using Docker and Docker Compose for reproducible local development and deployment readiness
Built a RAG system over CDC public health documents with PDF parsing, recursive chunking, FAISS vector retrieval, and hybrid search. Deployed on AWS with Streamlit frontend.
Tech: LangChain, OpenAI, FAISS, Streamlit, AWS
Agentic AI Workflows - Multi-Agent LangGraph
Designed multi-agent AI workflows using LangGraph with ReAct/ReWOO architectures, stateful pipelines, conditional routing, memory management, and human-in-the-loop checkpoints.
Tech: LangGraph, LangChain, OpenAI, Streamlit
Enterprise MCP Connector Hub
Built a Python-based MCP server using FastMCP for standardized, secure tool access for LLM agents. Enables agents to query Databricks and Snowflake without exposing credentials.
Built GenAI RAG platform enabling operations and finance teams with ~60% faster information retrieval across 50K+ documents. Deployed MCP-based connectors for secure agent-to-data access.
Manufacturing & Supply Chain (Carrier HVAC)
Delivered LLM-powered document analysis reducing manual review by ~40%. Ensemble forecasting models improved demand planning accuracy by 30-45%.
CRM & Sales Analytics (Sabnext Solutions)
Built gradient boosting models for churn prediction, lead scoring, and revenue forecasting used by sales leadership for quarterly planning.
Time-Series & Financial Analytics (Vassarlabs)
Built forecasting models (Prophet, ARIMA, LSTM) for supply chain and financial planning. Automated ML workflows with containerized pipelines on AWS.
Data Engineering & Automation (Vassarlabs India)
Built Python automation frameworks achieving 99% data accuracy, ETL pipelines, and reporting dashboards in Power BI and Tableau.
MLOps & Production Excellence
01
Experiment Tracking
MLflow for versioning, metrics logging, and model registry with comprehensive experiment management
02
Containerized Deployment
Docker + Docker Compose for reproducible builds; ECS with blue-green deployment patterns and autoscaling policies
03
CI/CD Pipelines
GitHub Actions for automated testing, building, and deployment; Amazon ECR for container image management
04
Production Monitoring
CloudWatch dashboards for real-time monitoring, alerting, and observability of AI services and APIs
05
Security & Access Control
IAM role-based authentication, API Gateway security controls, and MCP-based credential management for agents
06
Model Evaluation
RAGAS and DeepEval for LLM output evaluation; custom scoring rubrics and automated evaluation pipelines before production rollout
Established comprehensive MLOps practices supporting 15+ production models with consistent performance, automated monitoring, and rapid iteration cycles.
Data Engineering & Optimization
Pipeline Optimization
Optimized Databricks pipelines using Delta Lake, partitioning strategies, and query optimization. Built batch forecasting jobs with S3 artifact storage and scheduled weekly model refresh cycles.
Cloud Data Infrastructure
Deep AWS expertise: S3 for artifact storage and dataset versioning with IAM access controls. Migrated on-premise analytics workloads to AWS EC2 and S3 with lifecycle management policies.
Anomaly detection pipelines for CRM and supply chain data
Batch forecasting pipelines with weekly refresh cycles
Leadership & Collaboration
Cross-Functional Partnership
Worked cross-functionally with product, architecture, and business teams to operationalize AI across 10+ teams, translating complex requirements into scalable AI solutions
Enterprise AI Adoption
Drove successful adoption of GenAI tools across operations and finance teams at AUConnects, enabling 500+ users with self-service AI capabilities
Stakeholder Communication
Presented AI solutions and results to business leadership; built reporting dashboards in Power BI and Tableau for monthly performance reviews
Technical Mentorship
Collaborated with engineering teams on containerized deployment patterns, CI/CD best practices, and MLOps workflows across multiple organizations
Agile Delivery
Delivered production AI systems iteratively with version-controlled deployment workflows, A/B testing, and continuous monitoring for reliability
Let's Build the Future Together
I'm passionate about building production AI systems that solve real business problems. With 8+ years of experience in GenAI, RAG architectures, agentic workflows, and end-to-end ML pipelines, I bring both deep technical expertise and a track record of measurable impact.
Currently open to senior AI/ML Engineer roles focused on Generative AI, Agentic Systems, and Production ML — remote or relocation to Atlanta, GA.