Harshit Saxena
Data Scientist
Mumbai, Maharashtra, India
Summary
👤
SeniorHarshit Saxena is a Lead Data Scientist in Mumbai with 10 years of experience building production recommender systems and personalization at scale for Indian e-commerce and content platforms. He has driven end-to-end ML solutions at Meesho and Trell—from cold-start frameworks using CLIP and vector search to GBDT and neural rankers that improved retention and engagement across vernacular audiences. At Fractal he delivered revenue-impacting piracy detection models and automated ML pipelines, blending NLP, unsupervised attribution, and operational deployment with Jenkins/Oozie. Trained at IIT Madras, he combines strong research roots with hands-on engineering and a track record of running rigorous A/B experiments and mentoring teams. Notably, he has reduced time-to-view for new content through custom recommender math and integrated OpenAI-style embeddings into content seeding—showing a focus on practical, growth-oriented ML engineering.
11 years of coding experience
Github Skills (12)
medical-imaging10
jupyter-notebook10
deconvolution10
convolutional-neural-networks10
image-segmentation10
python9
tensorflow9
dicom9
animation7
video-editing7
video-processing6
gif5
Programming languages (3)
Github contributions (5)
harsh1795/CNN-DICOM-Segmentation
Jun 2017 - Feb 2018
DICOM Image Segmentation with CNNs in Tensorflow
Contributions:37 commits, 25 pushes, 1 comment in 8 months
dicomimage-segmentationtensorflowdicom-imagescnn
harsh1795/Hackathons
Feb 2017 - Jan 2021
Contributions:211 commits, 186 pushes, 3 branches in 3 years 11 months
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.