Hi, I'm Kiran Bodipati
AI & Machine Learning Engineer • Researcher
I am an AI and ML Engineer at HERE Technologies in Chicago, building production LLM and agentic workflows, chat interfaces, MCP servers, and Agentic BI tools.
Beyond engineering, you'll find me Scuba Diving, travelling, hiking, writing poetry, or exploring the outdoors.
About Me & Research
A glimpse into my academic foundations, industry engineering experience, and scientific interests.
Kiran Bodipati
HERE Technologies • Chicago, IL
Hi! I'm Kiran Bodipati, an AI and ML Engineer at HERE Technologies based in Chicago. I architect and engineer production-grade LLM and agentic workflows, conversational chat interfaces, Model Context Protocol (MCP) servers, and Agentic BI tools powered by Graph RAG and semantic search. My work also centers on robust MLOps and scalable cloud AI infrastructure, designing low-latency semantic caching systems with DynamoDB and Redis, knowledge graph querying, and containerized microservices. I completed my Master of Science in Data Science at the University of Michigan, Ann Arbor with a 4.0 GPA (Honored as Outstanding First Year Student and Valedictory Student Speaker), following my B.S. (Honours with Distinction) in Data Science & AI from Nanyang Technological University, Singapore.
Academic Honors
Excellence in Data Science & AI
M.S. in Data Science • Ann Arbor
B.S. in Data Science & AI • Minor in Business
Core Research & Exploration
Technical Focus & Research
Beyond The Screen
Current Role
AI & ML Engineer, HERE Technologies
Core Focus
Agentic AI & Predictive Analytics
Cloud & Infra
AWS, Terraform & Redis
Alma Mater
Univ. of Michigan (4.0 GPA) & NTU
bodipati@umich.edu
Interests
Scuba Diving, Hiking, Travel, Poetry
Experience & Education
A comprehensive record of my industry roles, academic degrees, peer-reviewed research, and demonstrated technical domains.
Work Experience
AI and ML Engineer
Full-timeHybridCurrent Role- Architecting and deploying production LLM and agentic workflows, conversational chat interfaces, Model Context Protocol (MCP) servers, and Agentic BI tools driven by semantic search and predictive analytics.
- Engineering MLOps and cloud AI infrastructure for agentic applications, including low-latency semantic caching using Amazon DynamoDB and Redis, knowledge graph querying, containerized microservices on AWS ECS, and serverless compute via AWS Lambda orchestrated with Terraform.
Data Science Intern
InternshipHybrid- Developed an automatic data profiling and quality tool leveraging statistical and LLM profiling to generate data-agnostic quality rules and ODCS-compliant data contracts.
- Developed an automatic distributed validation pipeline using Great Expectations and Prefect, improving runtime by 90%.
Co-Founder & Data Scientist
Co-FounderHybrid- Spearheading technology vision and engineering AI-driven real estate solutions; built and deployed an automated land analytics system to evaluate property zoning, pricing, and feasibility.
- Architected a real-time foot-traffic and customer behavior analytics platform using PyTorch facial identification to measure customer dwell time and visit frequency for commercial properties.
Data Engineer
Full-timeOn-site- Engineered distributed scraping and OCR pipelines using Python, Tesseract, and IndicTranslate to extract and parse over 3 million historical revenue court records and land deeds into Apache Solr.
- Built an automated Graph Neural Network document validation engine that predicts missing title chain links with 91% accuracy.
Data Science Intern
InternshipHybrid- Performed root cause analysis for manufacturing fault detection using Bayesian Belief Networks (BBN), achieving a 15% reduction in machine downtime.
- Engineered incremental learning preprocessing pipelines with comprehensive Pytest suites, and deployed interactive diagnostic dashboards using FastAPI, Streamlit, and Plotly.
Data Science Intern
InternshipHybrid- Built an end-to-end brand and logo recognition pipeline for e-commerce listings using OpenCV, YOLOv5, and PyTorch.
- Finetuned multi-class CNNs and engineered synthetic training edge cases, improving recall from 0.10 to 0.35 and precision to 0.60 across 150 brand classes.
Full Stack Project Intern
InternshipRemote- Designed an automated video game credits extraction engine using deep OCR to identify developer names and roles from game footage.
- Built a Flask and React web repository linking in-game credit profiles to verified developer LinkedIn accounts for real-time talent search.
Education & Academic Honors
Master of Science in Data Science
GPA: 4.0 / 4.0Maintained a perfect 4.0 / 4.0 GPA. Honored with the Outstanding First Year Student Award, and selected as the Valedictory Student Speaker for my cohort. Conducted research in LLM interpretability and multimodal vision-language architectures.
Bachelor of Science (Honours with Distinction) in Data Science & AI
Honours with DistinctionGraduated with Honours (Distinction) in Data Science and Artificial Intelligence, with a Minor in Business.
Research Papers & Publications
Benchmarking and Improving LLM Robustness for Personalized Generation
EMNLP 2025Chimaobi Okite, Naihao Deng, Kiran Bodipati, Huaidian Hou, Joyce Chai, Rada Mihalcea
Findings of EMNLP 2025 (Suzhou, China)
Text Restoration Using Image Super Resolution
ThesisKiran Bodipati (Advised by: Prof. Chen Change Loy)
Nanyang Technological University, Singapore
Explainability of Deep Learning-based Graph Embeddings
Conference PaperKiran Bodipati
The Student Conference on AI (STCAI 2021), Singapore
Statistical Analysis of the Effectiveness of Card-counting on Indian Rummy
Conference PaperVarun S Iyengar, Kiran Bodipati, Abhishek Vaidyanathan
The Student Conference on AI (STCAI 2021), Singapore
Core Technical Domains & Stack
Agentic AI, LLMs & Graph RAG
10 competenciesIn Practice:Applied at HERE Technologies, UMich AI Lab, and EMNLP 2025
AWS, Cloud & AI Infrastructure
10 competenciesIn Practice:Production AI Infra & Semantic Caching at HERE Technologies
Data Engineering, Quality & Orchestration
9 competenciesIn Practice:Applied at HERE Technologies, TEAL India, and HP Inc
Software Engineering, CI/CD & Development
9 competenciesIn Practice:Enterprise Systems, Production Repositories & UI Demos
Featured Projects & AI Systems
A curated collection of deep learning pipelines, multimodal AI experiments, geospatial systems, and algorithmic engines.
AI & VisionText Restoration using Image Super Resolution
Developed improvements to the state-of-the-art Real-ESRGAN architecture specifically for Scene Text Image Super Resolution, with synthetic dataset generation and higher-order degradation models.
NLP & AudioRedPILS: Political Polarity Search Engine
Semantic search engine indexing subreddits and comments on Indian political discourse, paired with sentiment polarity and topic analysis models.
Systems & AnalyticsAI Agents for Playing Indian Rummy
Designed intelligent AI agents simulating card-counting and game-theoretic strategies in Indian Rummy to mathematically prove statistically significant win-rate advantages.
AI & VisionPoem2Reel: AI Text-to-Video Engine
An AI-powered pipeline converting written poetry into engaging multimedia short reels by synchronizing generative imagery, audio, and visual pacing.
NLP & AudioSpeaker Segmented Speech Emotion Analyzer
Multi-speaker conversation pipeline that performs speaker diarization and classifies emotional transitions throughout dialogue recordings.
Systems & AnalyticsPublic Transport Mode Classifier
Trained predictive machine learning models on human mobility GPS time-series data to accurately identify public transport modes for urban planning.
AI & VisionVisual Kinship Recognition via Siamese Networks
Engineered dual-branch Siamese convolutional neural networks to predict facial kinship relationships between family portrait pairs.
Systems & AnalyticsiNTUition v7: Gamified Virtual Hackathon Platform
A gamified 2D interactive virtual platform built for hosting the iNTUition v7 hackathon, enabling participants, mentors, and sponsors to network live via avatars, attend workshops, and explore sponsor booths in real time.
NLP & AudioMusicMoods: Cross-Modal Music Recommender
Cross-modal sentiment-based music recommender system leveraging NLP sentiment analysis of user mood inputs to generate personalized playlists.