Pavan Kumar Kollipara
I'm Pavan Kumar Kollipara, a results-driven Machine Learning Engineer with 6+ years of experience designing, training, and deploying scalable ML and Generative AI systems on AWS, supporting enterprise-scale data…
About
I’m Pavan Kumar Kollipara, a results-driven Machine Learning Engineer with 6+ years of experience designing, training, and deploying scalable ML and Generative AI systems on AWS, supporting enterprise-scale data pipelines and intelligent automation across multiple domains. I’ve built 50+ distributed models for prediction, recommendation, and classification on multi-terabyte datasets, and I’m proficient with PyTorch, TensorFlow, XGBoost, and SageMaker. I’ve led MLOps automation using MLflow, Airflow, GitHub Actions, ArgoCD, and Spinnaker, automating production pipelines for training, deployment, and monitoring. I also work with LLMs and RAG pipelines using LangChain, LangFuse, and Hugging Face to enable retrieval-based responses and scalable contextual reasoning.
Recently I’ve focused on building retrieval workflows, agentic reasoning systems, and multi-GPU production workloads, while collaborating with cross-functional teams to translate research into reliable, scalable solutions. I enjoy mentoring teammates, driving practical ML adoption, and delivering value through end-to-end system design, monitoring, and continuous improvement.
- 6 years' experience
- Joined November 2025
Industry experience
- Software & Internet
- Healthcare
- Education
- Professional Services
More experts in Baltimore, United States
- AI Developers for hire in Baltimore, United States
- AI Engineers for hire in Baltimore, United States
- App Developers for hire in Baltimore, United States
- Data Scientists for hire in Baltimore, United States
- Developers for hire in Baltimore, United States
- Full Stack Developers for hire in Baltimore, United States
- Web Developers for hire in Baltimore, United States
Work experience· 7 roles
Developed and fine-tuned LLMs using Hugging Face, LangChain, and OpenAI APIs on AWS SageMaker and EMR, integrating PySpark ETL pipelines to improve sensor data inference on 1M+ records. Built and optimized RAG pipelines with LangChain, Pinecone, and LangGraph, analyzing 500GB+ of sensor logs for faster and more accurate contextual responses. Implemented MLOps workflows automating 200+ model runs for tracking, drift detection, and reproducibility. Deployed agentic AI frameworks leveraging LangGraph, MCP, and MLflow, automating data prep, tuning, and scheduling for 30K+ records per run on multi-GPU clusters.
Developed and deployed 8+ ML models for fraud detection, traffic classification, recommendations, and engraving using PyTorch, Scikit-learn, and AWS SageMaker, automating ETL in Airflow, Glue, and Kafka for 100K+ daily transactions. Built real-time traffic filtering models using supervised and anomaly detection, blocking 250K+ bot requests daily across 15 global storefronts with AWS Step Functions and Lambda workflows. Collaborated on CNN-based text-to-engraving model on multi-GPU clusters, processing 50K+ text inputs and improving accuracy across 5K+ SKUs. Streamlined MLOps pipelines with MLflow, Prometheus, and Grafana, enabling 250+ retraining cycles and versioned deployments via SageMaker Pipelines.
Developed and deployed 8+ ML models for fraud detection, traffic classification, recommendations, and engraving using PyTorch, Scikit-learn, and AWS SageMaker, automating ETL in Airflow, Glue, and Kafka for 100K+ daily transactions. Built real-time traffic filtering models using supervised and anomaly detection, blocking 250K+ bot requests daily across 15 global storefronts with AWS Step Functions and Lambda workflows. Collaborated on a CNN-based text-to-engraving model on multi-GPU (NVIDIA A100) clusters, processing 50K+ text inputs and improving engraving accuracy across 5K+ SKUs. Streamlined MLOps pipelines with MLflow, Prometheus, and Grafana, automating 250+ retraining cycles and enabling versioned deployments via SageMaker Pipelines. Optimized deployments using FastAPI, Flask, Istio, and Cilium eBPF, achieving zero downtime across 15+ Apple Store environments; built 5+ Go-based CLI tools for automation. Enhanced scalability with Terraform, Helm, and AWS SDK, cutting environment s
Education
Certifications and awards
Hire Pavan Kumar Kollipara today
Post a job and invite Pavan, or contact them directly. No hidden commission.