Prithivi Da
Founding CTO at Donkey Stereotype
Bangkok, Bangkok, Thailand
Summary
🤩
Rockstar🎓
Top SchoolPrithivi Da is a Founding CTO and independent AI/ML researcher with 11+ years of experience building production-grade data, retrieval and multimodal ML systems from Bangkok. He helps organisations cut through AI hype to deliver high-ROI, cost-effective solutions—specialising in dense, sparse and hybrid embeddings, rerankers for RAG, vision+language models and XAI for NLP. As creator of widely used open-source projects like Gramformer and Parrot_Paraphraser and author of prize-winning Hugging Face demos, he pairs hands-on engineering with research-grade contributions (including four USPTO inventions). He founded Donkey Stereotype to release permissive SPLADE models, a top ColBERT release and other retrievers that are already used by vector DB vendors, and currently focuses on contrastive pretraining and efficient multilingual retrievers. A practiced inventor and social entrepreneur, he blends platform and product thinking with first-principles writing on AI, drawing on deep experience in data engineering, platform strategy and microservices.
11 years of coding experience
12 years of employment as a software developer
Bachelor of Technology (B.Tech.) Information Technology, Bachelor of Technology (B.Tech.) Information Technology at St. Peter's Engineering College
French, English
Stackoverflow
73reputation
41kreached
3answers
3questions
Github Skills (19)
transformers10
python10
machine-learning10
autocorrect10
grammar10
tokenizer10
grammars10
nlp10
beam-search9
file-transfer6
hash6
combinations6
hive6
hdfs6
java6
Programming languages (8)
Github contributions (5)
PrithivirajDamodaran/Gramformer
Jul 2021 - Dec 2022
A framework for detecting, highlighting and correcting grammatical errors on natural language text. Created by Prithiviraj Damodaran. Open to pull requests and other forms of collaboration.
Role in this project:
ML Engineer
ML EngineerContributions:6 releases, 65 commits, 14 PRs in 1 year 6 months
Contributions summary:Prithivi focused on enhancing the core functionality of the Gramformer project, which is designed for grammar correction. Their contributions centered on refining the QE (Quality Estimation) estimator, improving the model's accuracy, and integrating it with various components. This involved modifying the core `gramformer.py` file, adjusting dependencies, and optimizing the correction process. The user also explored the use of beam search and handled authentication tokens.
grammaticalphoneticscollaborationspellinghighlighting
PrithivirajDamodaran/Parrot_Paraphraser
Apr 2021 - Dec 2022
A practical and feature-rich paraphrasing framework to augment human intents in text form to build robust NLU models for conversational engines. Created by Prithiviraj Damodaran. Open to pull requests and other forms of collaboration.
Role in this project:ML Engineer
ML EngineerContributions:1 release, 1 review, 196 commits in 1 year 8 months
Contributions summary:Prithivi primarily contributed to the development and refinement of the `parrot` paraphrasing framework. Their work involved modifications to the `filters.py` file, specifically focusing on the `Adequacy` and `Fluency` filters, and incorporating the use of pretrained models. They made changes to improve readability and accuracy by updating the filtering logic. Furthermore, the user enhanced the project's functionality by adding support for the use of authentication tokens and making optimizations to the demo file.
nlprasa-nlunluintentsconversational
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.