[Oct'20 - ...] Appointed as Moderator of Machine Intelligence Group for newly founded IITH Chapter, ACM
[Sep'19 - Oct'20] Founding Coordinator of Machine Learning Club "Epoch", part of Sci-Tech council under Gymkhana, IIT Hyderabad
Club Activities:
Promoting Machine Learning within IITH community
Organize hackathons and related activities at inter-hostel and college level
Accelerating Graph Neural Networks
Successfully quantized GNNs with BFP16 on CPU and TPU
Prepared workload for Intrusion Detection using GNNs
Enabling accelerated GNN on large datasets (e.g. OGBn-Papers (110M nodes))
Selected to join Data Platform Research Team, Artificial Intelligence Research Center, Japan
Areas of research:
Graph Neural Networks
Explainability
Advised SpoonShot, a food-tech startup, on application of Machine Learning to food and related information.
For details, contact: Sai Sreenivas Kodur, CTO, SpoonShot
Worked remotely for 2-months with SentiSum, London, UK on Aspect Based Sentiment Analysis (ABSA) by taking multi-label sequence classification perspective for aspect extraction. Introduced Transformer based models such as BERT and used Knowledge Distillation to decrease latency.
Responsibilities and Accomplishments:
Improved performance (~5%) of existing Aspect Extraction pipeline with Transformer-based models
Decreased latency (5.7x) using Soft Knowledge Distillation
Incorporated better logging and configuration handling mechanism to existing code-base
Automated Root Cause Detection from Log Files using Machine Learning and Finite State Automata
3 years and 3 months work experience on enterprise Backup and Archiving product "NetBackup" as application software developer in the applications team. Received Symantec Applause Award (Level 2) for the project “Securing the NetBackup Client Daemon (BPCD)” as best team and team player. Worked as “Beta Buddy”, guided customers to setup, facilitate problems to the proper teams and handling customer problems for NetBackup 7.7 release.
Responsibilities and Accomplishments:
Developed features for Enterprise Backup and Archiving Application "NetBackUp"
Secured server to client communication and created over 500 automated test cases
Developed NetBackup Granular Recovery Technology (GRT) for single item recovery
Developed "Delayed Cataloging" feature for Microsoft Exchange (Email server) to support single email restore
5+ years of experience working with Python 3.x.
4+ years of experience on Machine Learning libraries (scikit-learn, Gensim, Numpy, Pandas, NetworkX).
3+ years of experience with Deep Learning framework Pytorch and related libraries (such as Pytorch Lightning, Deep Graph Library, AllenNLP, TorchText, FastAI, Transformers).
Familiar with Git, Github, BitBucket.
Comfortable with academic writing in LaTeX, Overleaf
Awarded $500 ACM SIGWEB Student Travel Award to present my paper at WebSci 2020 conference.
International Travel Grant of ₹75,000/- (~$1000) by Microsoft Research Lab, India to present my paper titled "Class Specific TF-IDF Boosting for Short-text Classification" at SMERP workshop at The Web Conference (Formerly WWW) at Lyon, France.
2nd Runner Up with ₹25000/- (~$350) out of 24 teams at 2018 Aerospace Automation Challenge Hackathon conducted by Honeywell Aerospace.
Received Symantec Applause Award (Level 2) twice for the project "Securing the NetBackup Client Daemon (BPCD)" as best team player.
Participated in "HackU-2011", an Open source application software development competition organized by Yahoo Inc. at IIT Kharagpur and our group was among the 26 teams who demoed.