Puneet Ludu
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
ML Engineer (Tech Lead scope) @ Zillow (Zestimate) | Production ML · LLM Reliability · RAG & GenAI | 13+ yrs
I'm a machine learning engineer with 13+ years of experience building and deploying production ML and LLM systems across real estate, dating, and financial-data domains. I hold an MS in Computer Science from the University at Buffalo (SUNY) and a B.Tech. in Computer Science from JIIT, India. Currently at Zillow, I work on the Zestimate, operating at tech-lead scope: I led the proof-of-concept and core architecture for Listing IQ (an AI-powered CMA tool for agents), shipped the first LLM assistant deployed to Zillow customer care (RAG-grounded, with output validation to guard against hallucination), and built the Active Listing Comps engine from the ground up (3M+ listings, powering dashboards with 8M+ monthly views). I also led infrastructure modernization that saved $500K+ annually, and I'm currently prototyping multimodal and explainable (NAM-based) approaches for next-generation valuation. Earlier, I optimized subscription pricing at OkCupid (Match Group), driving a 6% revenue increase, and built ML systems at FactSet including speaker identification, private-company fact extraction from 1.6M websites, and document search/ranking. I've published several papers on ML and data analytics, and I maintain open-source projects like Lotion (2K+ GitHub stars). I'm always eager to learn new technologies and tackle hard problems at the intersection of ML systems and product. Feel free to connect to discuss opportunities or ideas.
University at Buffalo
Master of Science (M.S.), Computer Science
January 1, 2013 – January 1, 2015
Jaypee Institute Of Information Technology
Bachelor of Technology (BTech), Computer Science and Engineering
January 1, 2006 – January 1, 2010
Zillow
Senior Machine Learning Engineer
September 1, 2021 – Present
OkCupid
Senior Machine Learning Engineer
May 1, 2020 – September 1, 2021
New York City Metropolitan Area
FactSet
Senior Machine Learning Engineer
January 1, 2018 – May 1, 2020
FactSet
Machine Learning Engineer
September 1, 2016 – January 1, 2018
FactSet
Software Engineer
April 1, 2015 – September 1, 2016
Tata Research Development and Design Centre (TRDDC)
Research Engineer ( Machine Learning / Big Data )
July 1, 2011 – August 1, 2013
Noida, Uttar Pradesh, India
Tata Consultancy Services
Developer (Assistant System engineer)
December 1, 2010 – June 1, 2011
Thiruvananthapuram Taluk, India
Indian Farmers Fertiliser Coop Ltd.
Software Developer (Intern)
April 1, 2009 – June 1, 2009
Greater Lucknow Area
Jaypee Institute Of Information Technology
Undergraduate Teaching Assistant
January 1, 2009 – June 1, 2010
Noida, Uttar Pradesh, India
Jaypee Institute Of Information Technology
Program Coordinator (Faculty)
December 1, 2008 – December 1, 2008
Noida, Uttar Pradesh, India
Memodiction
September 1, 2015 – Present
Memodiction lets you learn and revise words to build your vocabulary intelligently. It runs on an innovative algorithm which will allow you to keep reading words and learns which words are more difficult than others for you. Those and similar words would appear more often for your revision, hence enabling you to learn and revise hard words naturally.
LangBase - Ontology based ingestion system
February 1, 2014 – May 1, 2014
This project aims at creating an interface for ingestion of data and integrating it using Apache Jena (RDF). The data Consists of 1TB of Scanned Images, small audio/video clips of endangered languages, related to linguistics project undergoing in Cameroon. The project is supported by the National Science Foundation.
Tweetiments
January 1, 2014 – Present
A simple tool to know the mood elicited by your last 20 tweets. I am currently expanding this project.
Quena - A Question-Answering System
November 1, 2013 – Present
Quena is a question answering system based on Wikipedia data, on which you can ask any type of question, and it would mostly give you the answer you were looking for. Few salient features of Quena: • We have indexed complete Wikipedia data using Apache Solr, which comes out to be 1.6 Mn pages after cleaning. • We have designed our own Query (Question) Analyzer using Stanford NLP. • Quena can process several queries like "Who was doctoral advisor of Einstein", "What do Larry Page do" or "Who wrote Pride and Prejudice". • We have implemented our own ranking algorithm, which would give rank articles according to the popularity of the article on that Topic. • Further we used Normalized N-Gram similarity for measuring the similarity between two words. • Quena can disambiguate meaning of attribute asked according to the entity. For e.g. "How big is Sahara" would give me area of Sahara, while "how big is Dwayne Johnson" would give his height. • Quena is highly scalable
Search Engine based on Wikipedia Data
September 1, 2013 – Present
Working on a search engine Indexing, based on Wikipedia dump. Implemented parsing and removal of wiki markups. Followed by Tokenization and Index construction.
Auto Wallpaper changer
December 1, 2010 – Present
A small tool to change wallpapers automatically in Lubuntu, using a shell script.
Clustering of songs according to mood (Sentiment Analysis of music)
August 1, 2009 – April 1, 2010
Music is one of the basic human needs for recreation and entertainment. As song files are digitalized now a days, and digital libraries are expanding continuously, which makes it difficult to recall a song. Thus need of a new classification system other than genre is very obvious and mood based classification system serves the purpose very well. In this paper we will present a well-defined architecture to classify songs into different mood-based categories, using audio content analysis, affective value of song lyrics to map a song onto a psychological-based emotion space and information from online sources. In audio content analysis we will use music features such as intensity, timbre and rhythm including their subfeatures to map music in a 2-Dimensional emotional space. In lyric based classification 1-Dimensional emotional space is used. Both the results are merged onto a 2-Dimensional emotional space, which will classify song into a particular mood category. Finally clusters of mood based song files are formed and arranged according to data acquired from various Internet sources.
Windows 7 Theme for IceWm
July 1, 2009 – Present
Created a theme for IceWm (a very lightweight window manager for Linux). Got 2522 downloads and appreciation from across the globe.
English to Hindi Translator and Transliterator
January 1, 2009 – Present
Transliterator: • Romadeva (Roman to Devanagari) transliteration tool developed in Gambas, converts phonetics of English to Hindi, resulting in Hindi text using an English keyboard. Translator: • This tool translates English text to Hindi with meaning and change in grammar using Hidden Markov Model. This tool was simple to use, and can contextually translate.
Josh - An Ubuntu re-spin for JIIT
July 1, 2008 – Present
An Ubuntu re-spin with many tools built from scratch for e.g. Aptilan is an apt-get like tool written in Java which can sync all debian packages from the LAN is helpful for installing or updating software when internet speed is slow. Josh was deployed in a lab of my University.
PSLFS – Virtual File system with a CLI
August 1, 2007 – Present
A virtual file system with a DOS like command line tool to show features of the file system: • Engineered complete file system from scratch in C, using K-ary data structure. • Tested successfully on pen-drives and external media drives for encrypting data.
Winsiu - A very lightweight Desktop environment
January 1, 2007 – Present
It acts like a small desktop environment with very low graphics and can perform various desktop activities in Unix graphically i.e. Without using commands such as Creating/Deleting/Editing/Navigating files and folders, compiling the C/C++ file etc.
Sequence Models
Coursera
June 24, 2026 – Present
Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization
Coursera
June 24, 2026 – Present
Convolutional Neural Networks
Coursera
June 24, 2026 – Present
Structuring Machine Learning Projects
Coursera
June 24, 2026 – Present
Cultural Fit Analysis
The candidate's project history shows a strong inclination towards independent, innovative, and research-oriented work, ranging from operating system re-spins to advanced NLP systems. This demonstrates intellectual curiosity and a proactive approach to learning and building. The transition from Software Engineer to Machine Learning Engineer and then Senior Machine Learning Engineer across multiple companies (FactSet, OkCupid, Zillow) indicates adaptability and a continuous drive for growth. While the target role is 'Data Analyst', the candidate's experience is heavily skewed towards Machine Learning Engineering, which is a more advanced and specialized field. This might indicate a potential mismatch in the specific day-to-day responsibilities of a typical Data Analyst role, which often focuses more on reporting, dashboarding, and SQL-based analysis rather than building and deploying ML models. However, the underlying data analysis and interpretation skills are undoubtedly present and highly developed.
Soft Skills & Operational Fit
The candidate's resume highlights leadership, mentorship, and the ability to establish coding standards and workflows, indicating strong operational fit and soft skills for a senior role. Project descriptions suggest problem-solving, innovation, and a results-oriented approach. The experience as a Teaching Assistant and Program Coordinator also points to strong communication and instructional abilities.