I'm Logesh Kumar Umapathi, a Lead ML Engineer specializing in NLP, Deep learning and Machine learning. I build intelligent applications & experiences from ground up.

Get in touch logeshkumaru@gmail.com


I'm currently a Lead Machine learning research engineer at Saama building NLP/Deep learning products to accelerate clinical trials and time to market of drugs. I am also a Mentor at Springboard, helping future Machine learning engineers and Data scientists in the weekends.

When I'm not in front of a computer screen, I'm probably playing badminton, photographing, or reading a book.

  • Python
  • Javascript
  • Scala
  • HTML
  • CSS
  • Pytorch
  • Transformers
  • Tensorflow
  • Sklearn
  • langchain
  • Apache Spark
  • AWS Sagemaker
  • API Dev & Management
  • Lambda
  • Apache Airflow
  • Docker
  • EC2
  • Bash
  • Git & Github
  • Chrome DevTools
  • Postman
Research Publications

A large-scale (194k), Multiple-Choice Question Answering (MCQA) dataset designed to address realworld medical entrance exam questions.

A 15.5B StarCoder is a language model (LM) trained on source code and natural language text

A 1.1B parameter models for Java, JavaScript, and Python code generation

Featured Articles

The intent of this article is to introduce you to a simple active learning method called ‘Uncertainty sampling with entropy’ and demonstrate its usefulness with an example.

An attempt to understand features and patterns learnt by a Fine-tuned BERT model

Article exploring visual intuition of bayes rule.

Article featured in Becominghuman.ai on using a language identifcation model.

Article briefing common mistakes that are to be avoided in a tensorflow model.

Article describing a workaround solution to return binary data from lambda service.


we will explore the latest breakthroughs and discuss how LLMs can be used to solve complex problems that require reasoning and planning. By unlocking these capabilities, LLMs can be used to build sophisticated chatbots, intelligent assistants, and other NLP applications-

Reasoning Large language models NLP Prompt engineering

This hack session would include the techniques that can be used to interpret the decisions of RNNs, LSTMs and Transformer models.

Transformers Deep learning NLP

A talk focussing on the realities of ML and its biases. Cutting through the hype

Machine learning Biases Adversarial

A talk and Demo on how to get started in ML and build a quick prototype.

Machine learning Deep learning Colab
Other Personal Projects

A framework for the evaluation of autoregressive code generation language models.

Pytorch code generation LLMs

A library to synthesize text datasets using Large Language Models (LLM)

Pytorch Transformers LLMs

A scaffolding for keras and tensorflow with some callbacks , metrics and logs inbuilt.

Python Keras Tensorflow

A Repo to elaborate on two of my many attempts at solving Quora Insincere Questions Classification competition

Python Kaggle Deep learning Tensorflow