Rohith Andanala | Sr. Machine Learning Engineer

Hi, my name is

Venkata Sai Rohith Andanala

Senior Machine Learning Engineer

I build Production-Ready AI Systems that Drive Business Impact, specializing in Scalable Predictive Modeling, MLOps, and Generative AI solutions across Healthcare and Enterprise IT.

Explore My Case Studies

Want to know more about me!

$2.3M

Annual Savings (Automated Claim Intake)

92%

Data Accuracy (Document Intelligence Pipelines)

3+ Years

End-to-End ML & MLOps Experience

Experience & Education

My professional journey and academic background in computer science and machine learning.

Work History

Jun 2024 – Present, USA

Software Engineer – AI/ML

The Cigna Group

Enabled automation of 60% of claim intake, generating $2.3M in annual cost savings.
Integrated GenAI (GPT/Bedrock) for clinical note summarization, reducing manual review by 55%.
Deployed containerized ML/GenAI microservices using Docker/Amazon ECS, achieving 99.7% uptime.

Apr 2022 – Jul 2023, Hyderabad, India

AI/ML Engineer

Infosys

Built ensemble regression models reducing pricing error (RMSE) by 27% (R²=0.87).
Automated model workflows via AWS SageMaker/Databricks, cutting refresh time by 35%.
Used MLflow for governance and Airflow for drift monitoring.

Sep 2021 – Mar 2022, Hyderabad, India

AI/ML Engineer INTERN

SoftAge Group

Developed ML models predicting IT job failures, reducing downtime incidents by 25%.
Used TensorFlow (LSTM/Isolation Forest) and MLflow.

Education

2023-2024

Master of Computer Science

Fairfield University, Connecticut, USA

2017-2021

Bachelor of Metallurgical and Materials Science

Mahatma Gandhi Institution of Technology, India

Technical Skills

My expertise spans across the entire machine learning lifecycle, from data engineering to deployment and beyond.

Cloud & MLOps

▸AWS (SageMaker, Bedrock, Lambda, S3)
▸GCP Vertex AI
▸Docker
▸Kubernetes
▸MLflow
▸CI/CD Pipelines (GitHub Actions, Jenkins)

Machine Learning & Deep Learning

▸TensorFlow, PyTorch, Scikit-learn, Keras
▸CNNs, RNNs, Transfer Learning
▸Feature Engineering
▸Model Evaluation & Optimization

Databases & Data Engineering

▸PostgreSQL, MySQL, MongoDB, ElasticSearch
▸Pinecone (Vector DB)
▸Apache Spark
▸ETL Pipelines

Programming & Scripting

▸Python, SQL, R

Data Visualization & BI Tools

▸Tableau, Power BI
▸Matplotlib, Seaborn, Plotly

Generative AI & LLMs

▸OpenAI GPT (3.5/4)
▸Hugging Face Transformers
▸LangChain, LlamaIndex
▸Prompt Engineering
▸Retrieval-Augmented Generation (RAG)
▸Fine-tuning & Embedding Models
▸Model Deployment using APIs

Natural Language Processing (NLP)

▸Text Classification
▸Named Entity Recognition (NER)
▸Sentiment Analysis, Summarization
▸Conversational AI
▸Document Parsing & Question Answering Systems

APIs & Integrations

▸RESTful APIs, OpenAI API
▸Hugging Face Inference API
▸FastAPI, Flask
▸Microservices Architecture

Collaboration & Development Tools

▸Git, GitHub, GitLab, Jira, Confluence
▸VS Code, Jupyter Notebook

Projects

A showcase of my hands-on technical projects, from end-to-end MLOps pipelines to generative AI models.

Fake News Detection Platform

MLOps End-to-End

Built a fully operational fake news detection system leveraging NLP. Developed and deployed classification models using Amazon SageMaker, automating data ingestion via AWS Lambda and EventBridge to fetch real-time headlines every three days. The inference pipeline was served through a FastAPI backend on EC2, with performance tracked in MLflow. The frontend, crafted in Vue.js and Bootstrap, is hosted on GitHub Pages.

AWS SageMaker

Lambda

EventBridge

EC2

FastAPI

MLflow

Vue.js

Bootstrap

GitHub Actions

Live Project GitHub

YOLOv5 Pothole Detection System

Computer Vision

Designed and deployed a computer vision system using YOLOv5 to detect potholes in road images. Trained the model on over 2000 custom-labeled images using RTX 4060 GPU acceleration. Included a Tkinter-based GUI for image upload and real-time bounding box predictions.

YOLOv5

PyTorch

OpenCV

Tkinter

NumPy

RTX 4060 GPU

GitHub

GAN for Handwritten Digit Generation

Generative AI

Implemented a Generative Adversarial Network (GAN) to generate synthetic handwritten digits using the MNIST dataset. The model was trained over 50,000 epochs, resulting in significantly improved image quality. Project showcases understanding of adversarial learning and generator-discriminator dynamics.

TensorFlow

Keras

MNIST

NumPy

Matplotlib

Medium Post

LSTM Stock Price Predictor

Time-Series & RNNs

Developed an LSTM-based recurrent neural network to predict Google stock prices using historical timestamp data. The model architecture utilized multiple LSTM layers with dropout regularization. Demonstrated effective time-series forecasting with a low Mean Squared Error (0.0015).

TensorFlow

Keras

NumPy

Pandas

Matplotlib

GitHub