RAGHUNATHAREDDY GR
AI/ML Systems Engineering · MLOps · Reinforcement Learning
Bengaluru, India · B.Tech, Artificial Intelligence & Machine Learning · REVA University
01 / About ◇ 02 / Systems ◇ 03 / Research ◇ 04 / Stack ◇ 05 / Connect
My focus is on the engineering around the model — making systems reproducible, observable, testable, and deployable rather than stopping at a training notebook. Focus What I build and explore MLOps / LLMOps Model lifecycle workflows, experiment tracking, drift monitoring, CI/CD, RAG and agentic applications Cloud-native ML Containerized inference, Kubernetes deployment, autoscaling patterns and telemetry Reinforcement learning PPO, SAC, TD3, continuous control, ablations and robustness experiments Data systems Feature engineering, feature stores, data validation and analytics pipelines
02 / Selected systems A curated route through the projects. Each card links to its source repository; implementation details and setup instructions live there.
|
Python SDK and CLI for managing ML deployments to edge devices, with offline-first operation and drift-aware workflows. Python · Edge ML · CLI · Model deployment |
Customer-churn ML lifecycle with orchestration, experiment tracking, champion/challenger patterns, explainability and monitoring. MLflow · Airflow · FastAPI · Docker · SHAP |
|
LLM-assisted code review with schema-validated analysis and a feedback loop for tracking accepted findings and false positives. TypeScript · LLMs · Zod · PostgreSQL |
Shape-classification inference service with a FastAPI API, Docker packaging, Kubernetes deployment, HPA and Prometheus/Grafana observability. PyTorch · FastAPI · Kubernetes · Prometheus |
|
Data-centric transaction feature engineering with point-in-time correctness and feature-store concepts. PySpark · Feast · DVC · Redis · AWS |
Real weather data, realtime inference, prediction logging, drift reporting, retraining workflows and operational dashboards. Python · FastAPI · MLflow · Evidently · Grafana |
More applied AI, analytics and systems projects
Project Focus Ticket Classification MLOps DistilBERT classification, MLflow registry, drift monitoring and CI/CD Churn + Retention Agent Churn prediction, feature management and an LLM-assisted retention workflow AS³ — Autonomous Synthetic Space Scientist Space telemetry, anomaly detection, multi-agent orchestration and visualization Fog-IDS Smart City Resource-aware multi-class intrusion detection for IoT/fog environments Fraud Detection on AWS ECS FastAPI inference, Docker and AWS ECS deployment Geospatial Measurement API Geospatial ingestion, coordinate transformations and geometric measurements Bluestock Mutual Fund Analytics Fund performance, investor analytics and dashboard reporting FinTech Internship Work Financial data extraction, analysis and software fundamentals Prism — Stellar Soroban Rust and Soroban smart-contract exploration for tokenized liabilities|
From-scratch continuous-control implementations and a reproducible comparison harness, with multi-seed experiments and statistical analysis. PyTorch · Gymnasium · MuJoCo |
Investigating how policy performance changes under domain and dynamics variation using domain-randomized reinforcement learning experiments. PPO · Domain randomization · MuJoCo |
PPO From Scratch repository · continuous-control demonstration
Build → Measure → Learn → Improve.