Devashish Deshpande - Staff Data Scientist

Devashish Deshpande

Staff Data Scientist

Bengaluru, Karnataka, India

Join Prog.AI to see contacts

Summary

👤

Senior

🎓

Top School

Devashish Deshpande is a Staff Data Scientist based in Bengaluru with 11 years of experience building and productionizing ML systems, particularly around real-time serving and agentic/knowledge-driven GenAI. He has led the end-to-end creation of Observe.AI’s KnowledgeAI RAG product and now focuses on agentic system benchmarking and evaluation at Talkdesk. His background blends strong research foundations (KU Leuven MSc in AI) with hands-on engineering—containerization, load-tested services, and deploying transformer-based models for tutoring and conversational NLU. An active open-source contributor, he has improved documentation and fixed core issues in high-profile projects like scikit-learn and matplotlib, and added topic-coherence tooling to Gensim. Outside work he channels the same systems-minded creativity into music composition and competitive team sports, reflecting a practical yet inventive approach to problem solving.

10 years of coding experience

7 years of employment as a software developer

Master's degree Artificial Intelligence, Master's degree Artificial Intelligence at KU Leuven

BITS Pilani, Birla Institute of Technology and Science

High School Science with computer science, High School Science with computer science at Delhi Public School - R. K. Puram

English, Hindi, Marathi, German, Dutch

Github Skills (19)

data-visualizations10

gensim10

python10

matplotlib10

scikit10

machine-learning10

data-visualisation10

topic-models10

scikit-learn10

topic-modeling10

data-visualization10

nlp10

documentation10

data-analysis10

data-science9

Programming languages (4)

TeXHTMLJupyter NotebookPython

Github contributions (5)

piskvorky/gensim

Apr 2016 - Nov 2016

Topic Modelling for Humans

Role in this project:

Data Scientist

Contributions:19 commits, 23 PRs, 144 comments in 7 months

Contributions summary:Devashish primarily focused on enhancing the Gensim library with improvements to topic coherence metrics and model converters. They implemented new coherence measures (c_uci, c_npmi), refined the input methods for topic coherence, and contributed to the conversion of models between different libraries, specifically adding functions to convert Mallet and Vowpal Wabbit models to Gensim's LdaModel. They also added tutorials and documentation to aid users in understanding and using these features.

pythonword-similarityword-embeddingsdata-miningfor-humans

scikit-learn/scikit-learn

Dec 2015 - Feb 2016

scikit-learn: machine learning in Python

Role in this project:

Data Scientist

Contributions:14 commits, 16 PRs, 150 comments in 2 months

Contributions summary:Devashish primarily contributed to improving documentation and fixing issues within the scikit-learn library, particularly related to datasets and cross-validation. They addressed documentation inconsistencies, such as fixing broken links and clarifying descriptions in the LFW dataset documentation, and also fixed a bug that arose in the StratifiedKFold implementation. Further work included updates to accommodate new functionality in the FeatureHasher class and corrections in the documentation for f_regression and DBSCAN. These actions demonstrate an emphasis on maintaining the library's quality and usability.

data-analysispythonstatisticsdata-sciencelearn-machine-learning

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.

Request Free Trial