🟢 Available for Opportunities

AI / ML Engineer

Prem Babu
Kanaparthi

I build production-ready AI systems, working across large language models, retrieval-augmented generation, and MLOps.

My focus is on scalable, reliable machine learning systems that perform in real-world environments. I'm currently pursuing a Master's in Artificial Intelligence at Rochester Institute of Technology.

Prem Babu Kanaparthi

About Me

A brief snapshot of how I think about building real-world AI systems.

Who I Am

I'm an AI/ML engineer focused on building production-ready systems across large language models, retrieval-augmented generation, and MLOps.

I care deeply about reliability, scalability, and shipping machine learning systems that work in real-world environments — not just experiments.

Rochester, NY
MS in Artificial Intelligence — Rochester Institute of Technology

What I Work On

Large Language Models & Retrieval-Augmented Generation
MLOps & Production Machine Learning Systems
Computer Vision & Applied Natural Language Processing

A snapshot of my experience so far

3+
Years of Hands-On ML Experience
10+
End-to-End Systems Built
36+
Tools & Frameworks Used
2+
Research / Technical Publications

Work Experience

A progression through real-world AI and ML roles, focused on production systems.

Generative AI Engineer

Concentrix + Webhelp•Newark, CA•Feb 2024 - Jul 2024
  • •Built multi-model LLM inference platform achieving 1.5s latency with 18% cost reduction across 25+ deployments
  • •Designed LiteLLM routing system with AWS Bedrock and SageMaker achieving 95%+ accuracy
  • •Implemented comprehensive guardrails and monitoring reducing incidents by 42%
  • •Built production evaluation pipelines for automated quality validation and model performance tracking

Data Science Intern

AlphaBits Technologies•India•Aug 2023 - Jan 2024
  • •Reduced search model iteration time by 90% and improved ranking accuracy by 10%
  • •Built scalable PySpark ETL pipelines and centralized feature store for data consistency
  • •Developed TF-IDF and Naive Bayes ranking models for production search systems
  • •Created ML Playbook adopted by 3 teams, reducing integration errors by 35%

ML Engineer Intern

iNeuron.ai•India•Jun 2023 - Aug 2023
  • •Built phishing detection system achieving 92% accuracy with 20% fewer false positives
  • •Implemented TensorFlow/Keras MLP with advanced optimization techniques
  • •Conducted F1-driven hyperparameter tuning for production deployment readiness
  • •Developed comprehensive testing and validation pipelines for model reliability

Featured Projects

Selected work demonstrating production AI systems and real-world deployment.

PaperMind - Autonomous arXiv Research Assistant

PaperMind - Autonomous arXiv Research Assistant

Multi-agent RAG system for literature reviews and research gaps

RAGLangChainMulti-AgentNLP
Click to view details

PaperMind - Autonomous arXiv Research Assistant

Multi-agent RAG system for literature reviews and research gaps

  • •Multi-agent RAG system for automated literature reviews and research gap analysis
  • •Specialized LLM agents with vector database coordination for efficient knowledge retrieval
  • •PDF ingestion and embedding pipelines for seamless document processing
  • •Intelligent query routing and synthesis across multiple research papers
View on GitHub
Real-Time Feature Store & ML Serving Platform

Real-Time Feature Store & ML Serving Platform

Streaming pipeline with Feast and Kafka: 5,000 events/hour, 45ms latency

MLOpsKubernetesReal-timeFeature Engineering
Click to view details

Real-Time Feature Store & ML Serving Platform

Streaming pipeline with Feast and Kafka: 5,000 events/hour, 45ms latency

  • •Built streaming pipeline with Feast and Kafka processing 5,000 events/hour with 45ms latency
  • •Deployed 3 production models: fraud detection (88% accuracy), recommendation (0.76 NDCG@5), forecasting (12% MAPE)
  • •Implemented MLflow monitoring with Kubernetes auto-retraining for model lifecycle management
  • •Designed scalable feature engineering pipelines with real-time and batch processing capabilities
View on GitHub
Multi-Agent Document Intelligence System

Multi-Agent Document Intelligence System

Automated text extraction: 82% accuracy, <5s processing

Multi-AgentRAGFastAPIDocument Processing
Click to view details

Multi-Agent Document Intelligence System

Automated text extraction: 82% accuracy, <5s processing

  • •Automated text extraction with 82% accuracy and sub-5-second processing time
  • •RAG implementation with Qdrant vector database: 500+ chunks, 85% top-3 relevance
  • •FastAPI production service reducing manual data entry by 84%
  • •Multi-agent orchestration for document classification, extraction, and validation
View on GitHub
Advanced Image Classification Pipeline

Advanced Image Classification Pipeline

Deep learning pipeline for image classification with 94% accuracy

Computer VisionTensorFlowCNNTransfer Learning
Click to view details

Advanced Image Classification Pipeline

Deep learning pipeline for image classification with 94% accuracy

  • •CNN-based image classification achieving 94% accuracy on custom dataset
  • •Transfer learning with ResNet and EfficientNet architectures
  • •Data augmentation pipeline for robust model training
  • •Deployed with TensorFlow Serving for real-time inference
View on GitHub
Multi-Language Sentiment Analysis System

Multi-Language Sentiment Analysis System

Transformer-based sentiment analysis supporting 10+ languages

NLPBERTTransformersFastAPI
Click to view details

Multi-Language Sentiment Analysis System

Transformer-based sentiment analysis supporting 10+ languages

  • •Fine-tuned BERT model for multi-language sentiment analysis
  • •Supports 10+ languages with 89% average accuracy
  • •REST API built with FastAPI for real-time sentiment scoring
  • •Integrated with streaming data pipeline for social media monitoring
View on GitHub

How I Build AI Systems

A practical breakdown of the layers behind my work.

Modeling & Intelligence

Training, adapting, and evaluating models for real-world performance.

PyTorchTensorFlowScikit-learnHugging FaceLarge Language ModelsRetrieval-Augmented GenerationComputer VisionNLPDeep Learning

Inference & Applications

Turning models into reliable, usable applications.

LangChainFastAPIPrompt EngineeringFeature EngineeringReal-time Inference

Infrastructure & MLOps

Deploying, scaling, and monitoring production AI systems.

AWS BedrockAWS SageMakerDockerKubernetesMLflowCI/CDGitHub ActionsTerraformPalantir Foundry

Data & Platforms

Building data pipelines and analytics foundations.

PySparkSQLSnowflakePostgreSQLMongoDBPandasNumPy

Education & Achievements

Academic background and research experience.

Education

Master of Science in Artificial Intelligence

Rochester Institute of Technology•Rochester, NY•Aug 2024 - May 2026

Bachelor of Technology in Computer Science

NIT Silchar•India•Aug 2020 - May 2024

Research

Lightweight Channel Attention for Efficient CNNs

Research on efficient attention mechanisms for CNNs

View paper

Certifications

Stanford Machine Learning Specialization

View certificate

Resume

View or download my professional resume.

Prem Babu Kanaparthi

AI/ML Engineer — MS in Artificial Intelligence

AI/ML engineer focused on production LLM systems, RAG pipelines, and scalable ML infrastructure.

Download PDFView Online

Last updated: January 2025

Get In Touch

I'm always open to discussing opportunities, collaborations, or interesting problems.