I'm an Agentic AI Engineer at Synchrony Financial with 3+ years of experience building autonomous multi-agent systems, production RAG pipelines, and scalable ML infrastructure. Currently pursuing my MS in Information Systems with a specialization in Big Data and AI/ML Engineering at Northeastern University.
What I'm working on:
- 🤖 Building LangGraph multi-agent orchestration systems processing 5,000+ monthly cases with 85% accuracy
- 🔄 Developing autonomous workflows with dynamic task coordination and real-time decision-making
- 📊 Architecting production RAG systems handling 20K+ documents with AWS OpenSearch
- 💰 Optimizing LLM costs - eliminated $180K annually by migrating from GPT-4 to Llama-3.1
- 🚀 Leading CrewAI proof-of-concepts for workflow automation
🤖 Agentic AI & LLM Engineering
🔍 RAG & Vector Databases
🧮 Machine Learning & Deep Learning
🔧 MLOps & Infrastructure
☁️ Cloud & Big Data
💾 Databases
LangChain | RAG | FastAPI | Neo4j | PyTorch
Production-grade RAG system for medical document intelligence:
- 🧠 Architected RAG pipeline converting multimodal medical documents into structured entities with 89% accuracy
- 📊 Built Neo4j knowledge graph vectorizing 15K+ medical concepts for semantic search
- ⚡ Deployed containerized FastAPI system handling 100+ concurrent requests
- 🔍 Implemented hybrid retrieval combining vector similarity and graph traversal
PyTorch FSDP | Flash Attention | ONNX | Quantization
Optimized ML training and inference infrastructure:
- 🚀 Built distributed training system achieving 90% GPU efficiency on dual-GPU setup
- ⚡ Boosted training speed 1.7× processing 158K multimodal samples using Flash Attention
- 📉 Reduced memory consumption by 16% through efficient attention mechanisms
- 🎯 Maximized inference achieving 5× throughput using ONNX export and INT8 quantization
XGBoost | Kafka | Spark | Kubernetes | MLOps
Real-time fraud detection system with complete MLOps pipeline:
- 🎯 Achieved 99.6% accuracy with XGBoost model using SMOTE for class imbalance
- 🔄 Built real-time processing pipeline using Kafka & Spark for streaming data
- 🚢 Implemented end-to-end MLOps with GitHub Actions, Argo CD & GKE
- ⏱️ Optimized for <100ms inference latency at 100K+ TPS scale
Kubernetes | Docker | Jenkins | GitOps
Production MLOps platform for streamlined model deployment:
- 📦 Containerized ML applications with Docker & Kubernetes orchestration
- 🔄 Automated CI/CD pipelines via Jenkins & GitHub Webhooks
- 🚀 GitOps-based deployments with version control & automated rollbacks
- 📊 Optimized data preprocessing reducing runtime by 66%
Sep 2024 - Present
Agentic AI & Multi-Agent Systems:
- Developed LangGraph multi-agent orchestration processing 5,000+ monthly cases with 85% extraction accuracy
- Implemented autonomous workflows with dynamic task coordination and contextual reasoning
- Led CrewAI proof-of-concept for retail partner integration automating merchant onboarding
Production RAG & LLM Optimization:
- Built production RAG system processing 20K+ unstructured financial documents monthly
- Eliminated $180K annual costs migrating sentiment analysis from GPT-4 to Llama-3.1
- Architected document intelligence pipeline achieving 85% extraction accuracy
ML Infrastructure & MLOps:
- Engineered PySpark/Airflow pipelines processing 200GB monthly with 35% latency reduction
- Reduced infrastructure costs 35% through intelligent autoscaling with Prometheus/Grafana
- Built CI/CD pipelines with Jenkins/GitHub Actions serving 20+ production models
Sep 2021 - Aug 2023
- Processed 10K+ unstructured documents with RoBERTa NLP model improving accuracy 23%
- Built automated classification pipelines achieving 92% accuracy for production workflows
- Deployed production models via Azure ML with containerized Docker workflows
Northeastern University - Boston, MA
Master of Science in Information Systems (Big Data & AI/ML Engineering)
Anurag University - Hyderabad, India
Bachelor of Technology in Electrical Engineering
- 🎖️ Oracle Cloud Infrastructure 2025 Certified Generative AI Professional
- 💻 Open Source Contributor: Ivy ML Framework - 20+ merged PRs optimizing matrix operations
- 🌟 Building autonomous AI systems that process millions of transactions monthly
