Mansi Mane
I am working as a Staff Machine Learning Scientist at Walmart and my interests include Machine Learning applications in Recommender Systems, Natural Language Processing and Computer Vision. Previously, I was working on LLM billion scale pre-training as Applied Scientist at Amazon. I did masters from Carnegie Mellon University focused on Machine Learning.
Work Experience
Staff Machine Learning Scientist, Walmart, Sunnyvale, USA
- Multi-objective recommendation systems across home page and item page. Achieved 34bps CTR lift and 11bps ATCs lift across 100M+ customers.
- Personalized Push notification recommendation system. Drove $17M+ GMV per week.
- LLM-powered campaign to product type mapping via RAG for 6K product types, reducing 2 weeks manual effort to an hour. 0.74 hit rate.
- Recipe generation pipeline using GPT and diffusion models, 71% title and 14% cooking step accuracy.
Applied Scientist II, Amazon (AWS), Santa Clara, USA
- Trained billion scale parameter NLP model from scratch with minimal loss in accuracy. To enable customers to train such models seamlessly on AWS, worked on SageMaker Model Parallel and HuggingFace integration which is being used by 30% of distributed training customers
- Researched different batch size scaling algorithms which reduced training time by upto 3 times for ResNet training.
- Developed deep learning approaches like Siamese and triplet network in TensorFlow by using multi-modal attributes like image and text data for complementary item recommendations.
- Worked on making PyTorch available in deep learning containers which is being used by 10000 users per week.
- Built deep learning infrastructure using the following AWS services: S3, EC2, ECR, SageMaker, CloudFormation, CloudWatch, CodeBuild, IAM.
Machine Learning Scientist, Walmart Labs, Sunnyvale, USA
- Deployed personalized item recommendation model (matrix factorization) serving 10M users, 8bps add-to-cart lift.
- Siamese network for complementary recommendations, 15bps CTR lift.
- Sequence-to-sequence machine translation for 20K product titles for voice assistants.
Research Internship, CyLab Biometrics Center, CMU, Pittsburgh, USA
- Weakly supervised object detection using AlexNet in PyTorch; 0.17mAP on PASCAL VOC.
- Pre-processed data for face parsing using Fully Convolutional Instance Aware Semantic Segmentation.
- Segregated medical images containing nucleus, cells and not containing those for pap-smear test by applying techniques like Gaussian smoothening, threshoding etc.
- Synthesized 2D face images at different poses by modifying existing code to project 3D fitted morphable model for face into 2D.
Publications
Campaign-2-PT-RAG: LLM-Guided Semantic Product Type Attribution for Scalable Campaign Ranking
Yiming Che, Mansi Mane, Keerthi Gopalakrishnan, Parisa Kaghazgaran, Murali Mohana Krishna Dandu, Archana Venkatachalapathy, Sinduja Subramaniam, Yokila Arora, Evren Korpeoglu, Sushant Kumar, Kannan Achan
Accepted at LLM & Agents for Recommendation Systems (LARS), WWW 2026
Product Title Generation for Conversational Systems using BERT
Mansi Ranjit Mane, Shashank Kedia, Aditya Mantha, Stephen Guo, Kannan Achan
Revised version accepted at The Web Conference (WWW) 2021
Complementary-Similarity Learning using Quadruplet Network (Code)
Mansi Ranjit Mane, Stephen Guo, Kannan Achan
Presented at Workshop on Recommender Systems in Fashion, ACM Recommender Systems (RecSys) 2019
Deep Learning based Head and Tail Localization of C. elegans (Code)
Mansi Ranjit Mane, Aniket Anand Deshmukh, Adam Iliff
Presented at ICML 2019 Workshop on Computational Biology
Workshop Organization
Workshop on Generative AI for E-Commerce
ACM Conference on Recommender Systems (RecSys) 2025, Prague, Czech Republic
Workshop on Generative AI for E-Commerce
ACM International Conference on Information and Knowledge Management (CIKM) 2024, Boise, Idaho
Other Machine Learning Projects
Zero Shot Learning for Image Classification: (Project Report) (Code)
2D GANs with Capsule Networks, CMU: (Project Report) (Code)
Weakly Supervised Object Detection and Localization: (Code)
Fast Super-Resolution CNN (FSRCNN): (Presentation) (Code)