Rajiv M- AI/ML Engineer

Major Projects

Restaurant Predictive Ratings System Using Social Media Insights

Design and development of an AI-powered platform that aggregates and analyzes restaurant data from social media channels. Led the backend development using Python and LLM frameworks, and integrated Amazon Bedrock for large language model operations. Focused on key performance metrics and customer behavior to generate predictive ratings and actionable summaries.

USA

Backend & Services:

Amazon Bedrock
LLM Application Framework
Python

Cloud & Database:

AWS Redshift
EKS
SQS

DevOps & Tools:

GitHub Actions
Serverless Framework
Unit test

Generative AI Tool for Student Behavior

Developed a customized generative AI solution that processes raw observation data submitted by users and transforms it into structured, client-ready reports. Utilized Langchain and OpenAI APIs to fine-tune prompt engineering for better report accuracy. Tailored to the full Python-based backend and integrated batch processing for scalable performance.

USA

AI/ML Frameworks:

LangChain
OpenAI API

Backend:

Python

Interactive Source Code Analysis Tool Using Generative AI

Engineered a developer-centric tool that enables interactive querying of source code through a chat interface. Integrated Langchain for natural language interactions and LLMs for providing code insights and library recommendations. Managed cloud infrastructure using Docker and AWS EKS, with CI/CD and testing frameworks like GitHub Actions and unittest ensuring reliability and maintainability.

USA

AI/ML Frameworks:

Langchain for Generative AI/ML Solutions

Cloud & Containerization:

AWS
Docker
EKS

DevOps & Tools:

GitHub Actions
Serverless Framework
Unit test

Backend:

Python

Audio Conversion and Generative AI Base Analysis Tools

Led the backend development of a solution that transcribes MP3 audio into English text using offline processing, meeting stringent privacy requirements. Applied Generative AI and speech processing techniques to analyze transcripts and accurately detect and tag speakers. Worked with Transformers, PyTorch, OpenAI Whispers, and WizardLM Engine to implement offline NLP pipelines in Python.

USA

AI/ML Frameworks:

Open AI whispers
PyTorch
Speech processing library
Transformers
WizardLM Engine

Backend:

Python

Scalable Social Media Analytics Platform for ROI Measurement

Developed an end-to-end analytics tool that helps clients measure the ROI of their marketing campaigns via social media insights. Created a compute-heavy, elastic pipeline using Apache Spark and Snowflake for time-series analysis. Containerized the application with Docker, deployed using AWS EKS, and managed DevOps with GitHub Actions and Serverless Framework.

USA

Data Processing:

Apache Spark
Pachyderm
Snowflake

Cloud & Containerization:

AWS
Docker
EKS

DevOps & Tools:

GitHub Actions
Serverless Framework
Unit test

Backend:

Python

Email Deliverability Monitoring and Optimization Solution

Built a digital mail processing system that prevents marketing emails from landing in spam folders. Designed mathematical models to simulate real-world mail transactions and monitor mailbox reputation. Developed backend services using NodeJS, AWS Lambda, and Aurora DB, and wrote Python scripts for automation and deliverability analysis.

USA

Backend & Services:

Aurora 
AWS Lambda
Node.js

Backend:

Python

DevOps & Tools:

GitHub Actions
Unit test

Framework:

Serverless Framework

AI-Powered Healthcare Recommendation Platform

Contributed to developing a healthcare solution that recommends nearby, affordable healthcare providers based on user symptoms. Implemented AI/ML modules using TensorFlow and Scikit-learn for symptom analysis and prediction. Managed data processing with Apache Spark and stored data securely in MySQL and AWS Elasticsearch. Ensured robustness and performance through Python-based microservices and GitHub Actions-based DevOps workflows.

USA