Aman Dalmia
Hi, I'm Aman. I am passionate about helping people through technology. I've recently started Plio, a non-profit with a mission to improve learning outcomes for rural children. 

In the past, I have worked as a Research Fellow at Wadhwani Institute for AI on early detection of pest infestation, identifying malnutrition in neonates using a smartphone video and developing a triaging tool for COVID-19 using cough sounds alone. I received my Bachelor's degree in Electronics & Communication Engineering from the Indian Institute of Technology Guwahati, during which I worked on using machine learning in several application domains like satellite imagery for detecting terrorist camps (HyperVerge), early detection of breast cancer through thermal imagery (Niramai) and finding new relevant patterns in the human genome for protein synthesis. Some of my past work has been published in journals on quantitative thermography and computational biology along with established AI conferences like KDD and various publishing platforms like the PyTorch blog and Weights & Biases. I also co-founded a start-up called Scholar Mate for a short while during my sophomore year where we tried to bring parents, teachers and students on a single platform.

I am an AI enthusiast and want to improve the way humans interact with technology. With an entrepreneurial mindset, I am constantly on the lookout for solutions to existent problems in our everyday life and look forward to creating something that helps to improve the lives of billions of people across the world. Specifically, I'm really interested in thinking about using AI in the real-world and studying various ethical dimensions of AI like learning with less data, bias in AI, adversarial robustness, model compression and explainability.

In my spare time, I love playing the guitar, hearing stories and am a cricket fanboy. My favourite habit is reading books across several topics. I love writing and am extremely passionate about education. I believe in the concept of lifelong learning and I am the co-founder of the AI club at IIT Guwahati, IITG.ai. I can occasionally be found blogging on Medium.

Twitter: @dalmiaman
​Github: @dalmia / @dalmiaman
News
  • I've co-founded, Plio - a new non-profit with an aim to improve learning outcomes for rural children through technology.
  • Top 5% (silver medal) in the SIIM-ISIC Melanoma Classification challenge on Kaggle (team name: themlguy) [blog][code] to be released soon].
  • Our paper titled "Pest management in cotton farms: an AI-system case study from the global South" has been accepted as a full conference paper at KDD 2020.
  • Member of the Program Committee for the workshop on "Machine Learning in Real Life" at ICLR 2020.
  • Attended NeurIPS 2019 held in Vancouver, Canada.
  • Attended the Tech Sprint Week (part of Google AI Impact Challenge) representing Wadhwani AI in London.
  • Journal Paper acceptance, Jun'19: Our paper titled "Exploring deep learning networks for tumour segmentation in infrared images" got published in the Quantitative InfraRed Thermography Journal. This was done during my internship at Niramai.
  • Attended the Google AI Impact Challenge Bootcamp representing Wadhwani AI in San Francisco. Here's a blog post describing my experience.
  • Won a $2 million grant from Google, May'19: My project is going to received a funding of $2M from Google, being one out of only 20 grantees across the world, and the only team from India, as part of the Google AI Impact Challenge.
  • Selected for the Udacity KPIT Autonomous Tech Scholarship to the Self-Driving Car Nanodegree program.
  • ​Received offers from Fractal Analytics, Ola, Wadhwani Institute of AI, Umbo CV (Taiwan) and Niramai.
  • Conference Paper acceptance, Mar'18 Berlin: Our paper on tumour segmentation from thermographic images as part of my internship at Niramai got published in the QIRT conference.
  • ICNDE, March '18: Among the 100 people selected from all over India to attend the International Conference on Networked Digital Earth to be conducted by IIT Kharagpur.
  • 1st place at Hack in the East, Jan '18: Created a prototype to demonstrate the idea of Waste disposal using Deep Learning on Drone Images and won a cash prize of Rs. 10,000.
  • 9th place in GS Quantify (among 270+ teams) all over India, Oct '17: The challenge involved 3 questions from Computer Science, Machine Learning & Quant. I was responsible primarily for the ML question and contributed partly to the Quant one as well. 
  • Received the international fellowship by fast.ai, Oct '17
  • MIT Media Lab Workshop, Jan '17: Among the 9 candidates who were selected from all over India to attend the workshop in Mumbai.
  • Received a scholarship for Udacity’s Android Developer Nanodegree Program from Google & Tata Trusts.
  • 2nd place in a B-plan competition conducted by Startup Assam, among 34 other teams in Guwahati.
  • 1st place in CodeIO, IIT Guwahati, Jan '16: Developed an android app (Eventify) along with a browser extension that directly sends specific details (chosen by the user) of the institute mail to the app using Firebase Cloud Messaging.
Open Source 
I am a strong advocate of open-source, and believe in Learning by Doing. I frequently implement various deep learning architectures, and I have recently started uploading my notes for various online courses along with the implementation of the algorithms discussed. In the past, I've contributed to scikit-learn, one of the most widely used machine learning library. I've also been maintaining chapter-wise summary of the Deep Learning book here. Feel free to have a look and if you feel that you want to contribute, do let me know.
My contributions can be seen here. I worked on resolving issues involving documentation, API changes, bug fixes & memory enhancements and reviewed pull requests by other contributors.