Rohith Srinivasa

Python, SQL, JavaScript, Bash, Java, C++, PostgreSQL, MongoDB, Redis, Scikit-learn, TensorFlow, Keras, Hugging Face, XGBoost, Spark, Airflow, MLflow, PowerBI, Docker, FastAPI, Kubernetes, Terraform, NGINX, Azure AKS, AWS, GCP

LLM Based Scam Detection using Llama-3

LLaMA-3-8B, LoRA, PyTorch, AWS EC2, Docker, NGINX, MongoDB, Selenium, Fine-tuned LLaMA-3-8B using LoRA on 18,000+ phishing site samples for text classification; achieved 98.85% accuracy on real-world detection tasks., Engineered a scalable, real-time inference pipeline integrating Selenium web scraping, MongoDB storage, and containerized microservices on AWS EC2., Operationalized containerized inference services using Docker and NGINX, handling over 15K daily requests with sub-250ms latency in high-throughput environments. 

Cold-Start Movie Recommender using LLM Embeddings

OpenAI Embeddings, Scikit-learn, FastAPI, React.js, Cosine Similarity, Engineered a hybrid recommender system using OpenAI embeddings and collaborative filtering, improving cold-start user coverage by 45% and boosting Precision@10 to 88% on MovieLens-100K., Achieved 88% Precision@10 on the MovieLens 100k dataset; optimized using cosine similarity re-ranking., Deployed the model as a REST API using FastAPI and integrated a React.js frontend, achieving 95% UI responsiveness under 300ms and supporting 100+ concurrent users. 

MLOps Pipeline for Bike Demand Forecasting

Azure AKS, MLflow, Apache Spark, XGBoost, Terraform, Prometheus, Grafana, Implemented a full MLOps pipeline with Spark + XGBoost, integrated with MLflow, and provisioned CI/CD infra via Terraform on Azure AKS., Improved forecast accuracy by 18% through weekly retraining and drift-based model updates.

Azure AI Fundamentals (AI-900), Google Data Analytics, HuggingFace Transformers, Advanced ML (ASU), NLP with BERT, Databricks Partner Accreditation: Data Engineering & ML Foundations - focused on scalable pipelines, Delta Lake, and ML deployment in production, IBM Data Science Professional Certificate (In Progress)