Recent Public Work
Participated in WSDM Cup - Multilingual Chatbot Arena [got a solo silver medal placing 31st out of 950 teams]: challenge was to develop a reward model (used in RLHF stage) for multilingual human conversations on the chatbot arena (formerly LMSYS). Finetuned LLMs as reward models in classification setting and used various techniques like multi-stage training (pretraining, finetuning), pseudo labelling, LoRA, QLoRA, efficient inference techniques, knowledge distillation, etc. [Code here]
Talks
I was a speaker at inaugural Weights & Biases MLOps conference (Fully Connected 2023). You can listen to my talk here. [Announcement LinkedIn Post]
🏆 Achievements
Click on the competition name to go to it's overview page and click solution + code links to see the approach and source code
| Competition | Placement | Organization | Code/Solution |
|---|---|---|---|
| DataSolve 2022 | 1st place | Wolters Kluwer | code |
| GIA Winner 2024 | CEO's choice award winner | Wolters Kluwer | |
| Amazon ML Challenge 2021 | 11th place (among 3200+ teams) | Amazon | solution + code |
| WSDM Cup - Multilingual Chatbot Arena | Silver medal (31/950) | Kaggle | code |
| U.S. Patent Phrase to Phrase Matching | Top 1%, silver medal (31/1889) | Kaggle | solution code |
| Bristol-Myers Squibb – Molecular Translation | Silver medal (50/874) | Kaggle | solution + code |
| Sartorius - Cell Instance Segmentation | Bronze medal (117/1505) | Kaggle | code |
| Happywhale - Whale and Dolphin Identification | Bronze medal (132/1588) | Kaggle | - |
| Kaggle - LLM Science Exam | Silver medal (123/2664) | Kaggle | - |
📝 Open-Source Contributions
- Fixed an example script on HuggingFace 🤗 transformers repository for XLA devices - [PR]
- Made the experiment trackers to launch only on main process in distributed setups on 🤗 Accelerate library - [PR]
- Fixed several examples and removed the check for main process... on 🤗 Accelerate library - [PR]
- Update several 🤗 transformers
no_trainerscripts leveraging 🤗 Accelerate... - [PR] - Contributed a report to Weights & Biases showcasing the integration of MONAI and W&B - [PR]
🛠 My Projects
Click on the project name to directly go to it's GitHub Repository and click demo app to see a live demo of the project
- QA Bot for Gradient Dissent podcast hosted by Weights & Biases: A Question-Answering bot built on top of OPENAI's LLMs and LangChain... [demo app]
- Text to Image Synthesis using Attentional GANs: A PyTorch re-implementation of the paper AttnGAN... [demo app]
- Text Summarizer: A transformer based extractive and abstractive text summarizer... [demo app]
- Healthify: a ML based website which predicts the disease based on symptoms... [demo app]
- GPU vs TPU comparison for computer vision: compared GPU and TPU for computer vision applications... [results]
- HARVESTIFY: a ML and DL based website which recommends the best crop to grow... [demo app]
- JARVIS: a simple voice assistant made using Python [demo video]
Blogs I've written on other platforms
- Using Google’s Agent Development Kit and Agent2Agent – building modular, interoperable multi-agent systems
- Deep Reinforcement Learning – how neural networks power learning in complex environments
- Supervised vs Deep vs Reinforcement Learning – comparing three core ML paradigms
- RAG techniques: From naive to advanced
- LLMs are machine learning classifiers
- Building a Q&A Bot for Weights & Biases' Gradient Dissent Podcast using LangChain and OpenAI
- DeepMind Flamingo: A Visual Language Model for Few-Shot Learning
- BLIP-2: A new Visual Language Model by Salesforce
- SetFit: Efficient Few-Shot Learning Without Prompts
- Is PyTorch 2.0 Faster Than PyTorch 1.13?
- 3D Segmentation with MONAI and PyTorch Supercharged by Weights & Biases
- HuggingFace is all you need for NLP and beyond
- Hugging Face Accelerate Super Charged With Weights & Biases
- Managing and Tracking ML Experiments