Sep 2022 – Jan 2023 · 4 mos

• Led development of MVP of the company from scratch, created data pipelines and backend microservices
• Designed and implemented the campaign creation module, earning revenue of more than 5k USD to date
• Collaborated across teams for the successful implementation of modules like authentication and admin dashboard

May 2022 – Jul 2022 · 2 mos · Bengaluru, Karnataka, India

• Integrated ranking and semantic features of search dataset, and trained various models like lgbm, M5, etc
• Learnt Big Data handling through PySpark, and trained ranking models on datasets over 100 GB
• Outperformed the benchmark for Irrelevant Classification and achieved a gain of around 100 bps
• Received Pre-Placement Offer based on the performance during the internship

Apr 2022 – May 2023 · 1 yr 1 mo

• Spearheaded a 3-tier team to conceptualize a 3-day event involving 22 IITs and reaching 40,000+ students & professors across IITs
• Procured a budget of INR 3.5 million through marketing & sponsorship to conduct 15 tech competitions for 2000+ participants
• Liaised with 50+ organizations to develop 13 tech problem statements in 20 domains of science, technology & entrepreneurship

Sep 2021 – Nov 2021 · 2 mos · India

• Established baselines for the task of Leakage Generation, using generative models such as BART, T5 and GPT2
• Implemented conditional generation using stable-baselines3, experimenting with policies such as A2C along with the state representations and topk methods
• Integrated custom Policy and Value Networks in conditional generation, and implemented Hierarchical RL enhanced by a custom reward network

Jun 2021 – Sep 2021 · 3 mos

• Climate Hedging Project
• ◦ Implemented a research paper on Hedging against Climate Change, creating the 2 indices mentioned in the paper
• ◦ Constructed vocabulary from 19 White Papers, implementing noun phrase chunking, contractions fixing, etc
• ◦ Used tf-idf vectorizer for constructing indices, using data from over 30 Twitter Handles and 10 news sources
• ◦ The aggregate index achieved a significance correlation value of 43% on backtesting
• Thematic Tagging Project
• ◦ Created a pipeline to automatically tag funds, using various Transformer models and statistical methods
• ◦ Implemented weighted vocabulary, keyword search, etc and experimented with various sentence transformer models.
• ◦ Achieved 93% precision for the theme Robotics, 89% for AI, etc.
• Backtesting and expansion of Tagging Strategies
• ◦ Worked on the extension of the tagging strategies to Alexandria Headlines, using AWS Athena and USE Transformers
• ◦ Implemented various filters like relevance, novelty,etc and also tested penalizing factor approach for levelling the scores
• ◦ The Portfolio outperformed the benchmark(S&P 500) by over 30%

May 2021 – Aug 2021 · 3 mos

• Trained and deployed ASR models for low resource vernacular languages like Hindi, Marathi, etc
• Used the MUCS 2021 dataset for Hindi to fine-tune the Citrinet-512 model from the NVIDIA Nemo library
• Deployed the trained model for inference using Triton server and NVIDIA Jarvis

Mar 2021 – Mar 2022 · 1 yr

• Hosted competitions for over 400 registered participants of Software Corner and ECDC, Techkriti’22
• Ideated and supervised events like IOPC, Speed Quant and ML Hackathon having a total budget of over INR 75k

Jan 2020 – Mar 2021 · 1 yr 2 mos

• Responsible for building the Website and Mobile App for Techkriti'20, an annual four-day inter-collegiate technical and entrepreneurship festival at IIT Kanpur

Mar 2021 – May 2021 · 2 mos

• Developed an extensive Notifier Application for handling AWS and ETL job failures using Slack, Twilio and Google Scripts
• Developed a Customer Attrition Model using decision tree algorithms to predict the churn of customers
• Implemented Multiple Locality Search Algorithm, and enhanced the Address Prediction Model

Jan 2021 – Feb 2021 · 1 mo

• Developed a 2-stage pipeline of Barcode Detection and Restoration, to reinstate the damaged barcodes accurately
• Used Tensorflow to train a ssd resnet‑50 model for barcode segmentation, achieving an mAP value of 0.845
• Utilized Pytorch to Train GANs using the Image Inpainting Method, to restore the segmented barcodes from earlier stage

Dec 2020 – Jan 2021 · 1 mo

• Worked on creating an end-to-end Medical chatbot based on RASA and the open source datasets on diseases,
• scraping the sites using BeautifulSoup
• Trained lgbm on GA data, for customer based coupon recommendation, and deployed it on AWS Lambda
• Used Selenium to create click-emulator bots, and used AWS Transcribe API to transcribe video calls

Jul 2020 – Aug 2020 · 1 mo

• Created an NLP model to detect PII Markers from Text,PDF and image, using Named Entity Recognition with AWS Comprehend,spaCy and Stanford NER Tagger
• Used Regex for pattern matching,Pytesseract and OpenCV for image redaction, PyPDF2 for PDF Redaction and Docker for SaaS model
• Achieved F1 score of 0.82 for Geographical,0.74 for Name and 0.59 for Organisation Name Parsers

Mar 2020 – Mar 2021 · 1 yr

• Built the web app for Life@IITK, pertaining to the daily needs of the 8000+ students in the campus
• Responsible for managing the website and servers of the Science and Technology Council, IIT Kanpur

Mar 2020 – Mar 2021 · 1 yr

• Expanded the reach of the club, worked on the club handbook and conducted orientation sessions
• Developed a wireless attendance management system as part of the internal hackathon, and worked on the MECAP‑Virtual Reality Summer Project.

Dec 2019 – Mar 2020 · 3 mos

• • Learnt full stack web development, and worked on building the ratification portal for the council

Feb 2020 – Mar 2021 · 1 yr 1 mo · Kanpur, Uttar Pradesh, India

• Created Career Portal and Resources Portal for the council, with over 5000 registered students, to apply for internship opportunities and access relevant resources
• Developed and managed the website for the Students’ Research Convention, a flagship event of the council, and the Office of International Relations, IIT Kanpur