Welcome to my GitHub!
I'm a Machine Learning Engineer based in New York, NY, with over 11 years of experience building robust AI and data-driven systems that power products at scale. I thrive on delivering high-impact solutions, mentoring teams, and contributing to open source.

• 🧑‍💻Current Role: Machine Learning Engineer atZillow (Zestimate Team)
• 📍Location: NY/NJ
• 🎓Education: MS in Computer Science (SUNY Buffalo), B.Tech (JIIT, India)
• 💡Mission: Architecting intelligent, scalable ML platforms and products that solve real-world problems

Languages:
PythonJavaC/C++BashJavaScriptHTMLSQL

Frameworks & Tools:
PyTorchTensorFlowPySparkFastAPIDjango
KubeFlowMetaflowDataBricksDockerKubernetes
TerraformGitlab CIMongoDBDocumentDB

• Realtime Valuation Platform: Designed and led an interactive CMA with property embeddings and comp APIs, boosting agent/buyer engagement and unlocking new revenue streams.
• ML Infra Modernization: Modernized valuation pipelines with Python, AWS, Docker, Terraform, saving $500k+ annually.
• Team Impact: Established ML coding standards, mentored new hires, and reduced on-call alerts by 95%.

• Revenue Optimization: Built ML pipelines to optimize subscription discounts, increasing revenue by 6% via A/B testing.
• Full-cycle ownership: Feature engineering, modeling, deployment, and monitoring.

• NLP & CV Projects: Speaker identification for earnings calls, fuzzy doc deduplication, company fact extraction, query expansion, real-time formula ranking.
• Big Data: Built scalable ETL and search features for financial data products.

• Pattern Mining: Time-series event detection, data harmonization frameworks for large enterprise data.

• Lotion: Unofficial Notion.so desktop app for Linux (2K+ ⭐, 60K+ downloads)
• Romadeva: Roman-to-Devanagari script converter (used by Translators Without Borders)
• jTextBrew: Java library for fuzzy string matching
• Quena: QA system indexing 1.6M Wikipedia docs

• Google Scholar
  • “Inferring Latent Attributes of an Indian Twitter user…” (ACM Hypertext 2015)
  • “Inferring gender of a Twitter user…” (CORR 2014)
  • “Architecture for Automated Tagging and Clustering of Song Files…” (IJCSI 2010)

✉️Email:puneet.ludu@gmail.com

• Organizer @MUFin Workshop (AAAI 2023, PKDD 2022): Fostering research at the intersection of ML and finance.
• Always open to collaboration or a chat about ML, data, or new tech!

“Code is like humor. When you have to explain it, it’s bad.” — Cory House

PinnedLoading

1. lotionlotionPublic

   Unofficial Notion.so app for Linux

   JavaScript 2.1k 89

2. reasonitreasonitPublic

   Python 5

3. RomadevaRomadevaPublic

   It is a simple tool to convert roman script to indic(Devanagari) script. As most Keyboards are English and to write in Indic script is difficult. It is easy to write Hindi in roman script this give…

   C 13 1

4. HappyOrSadHappyOrSadPublic

   Sentiment Analysis of a file (Script to know if your text's mood is Happy or Sad)

   Perl 5

5. davydany/sultandavydany/sultanPublic

   Sultan: Command and Rule over your Shell

   Python 676 35

6. rpy2/rpy2rpy2/rpy2Public

   Interface to use R from Python

   Python 667 81