Work
Experiences

• July, 2022 — Present
• TrustYou GmbH, Munich, Germany
• Working Student: Data Engineer

Responsibilities:

• Developing new matching tool to find the discrepencies in clients' files

• Working on optimizing the air-flow matching process

• May, 2022 — Present
• TUM.ai, Germany
• Active Member: Software Engineer

Responsibilities:

• Developed the Industry Phase application

• Developing TUM.ai Makethon website

• Nov, 2019 — May, 2021
• Adobe Systems Pvt. Ltd., India
• Senior Software Engineer

Responsibilities:

• Developed closed captions RESTful interface to assist participants who are deaf or hard of hearing to participate in meetings & applied extractive text summarization

• Implemented a Single sign-on feature in Adobe Connect to provide a unified login service

• Created Python scripts for build checking; decreased deployment time by 15%

• Developed a tool to measure audio-video latency in Adobe Connect

• July, 2018 — Oct, 2019
• Noida Power Company Limited, India
• Software Developer

Responsibilities:

• Led an initiative to develop the first-ever in-house smart conversational assistant “NPCL Saathi” employing RASA Core/NLU; reduced 11% footfall in the customer care offices

• Created a statistical model to analyze Theft Detection in consumption pattern

• Contributed to the development of the KRA (Key Result Areas) Management system for real-time tasks update

• Developed a tool to measure audio-video latency in Adobe Connect

• June, 2017 — July, 2017
• Delhi Metro Rail Corporation (DMRC, Govt. of India), India
• Summer Intern

Responsibility:

• Created a live project (website) for tracking work progress of DMRC workers at Stations and Depots; allowed real-time monitoring of workers and their tasks

Some Things I've Built

• Earthencity: Smart sustainable app
• Tech Stack: Python, Figma
Prototyped the smart sustainable food scanner app with a team of 5 in Science Hack by TUM:Junge Akademie, to help users to know the details of food items so that they can contribute to the nature. App shows Nutri score and sustainabality score so that user can take decision on what to buy. There is a Sustainable point system as well to motivate user in healthy buying.
View Pitch

• Human vs Computer Evaluation of Abstractive Summarization
• Tech Stack: Python, HuggingFace, Transformers, NLP
Used BART, Pegasus models to evaluate summarizarion on SAMSUM, DialogSUM, CRD3 datasets along with human evaluation. Used HuggingFace to fetch the dataset and model and then fine-tuned to summarize the given input. Then applied ROUGE metric to see the evaluation score.

• Finderbuddy— A procurement Bot
• Tech Stack: Python, Flask, SQLite
Chatbot to automate and ease the process of procurement activities in large firms; accepts the input in text and speech and provides output accordingly. Got selected in UnitedHCL Manchester Hackathon after clearing the 1395 participants as a finalist.

• People Wealth— Alexa Skill
• Tech stack: Python, AWS Lambda
The skill developed for Alexa that finds the wealth of celebrities. https://www.amazon.in/amanvars-PeopleWealth/dp/B07H4SH176

• Search Station— Chrome Extension
• Tech stack: JavaScript, HTML
The extension helps users search across famous indian e commerce websites like Flipkart, Amazon among others

• ZHCET TPO— Android App
• Tech stack: Java, SQLite
Notification app that provides students with day-to-day updates regarding job placements and other essential information on the go

• AO Bot— A conversational dialog engine
• Tech stack: Python, MongoDB, Flask
Dialog system that interacted with the user in textual format and supported Hinglish input (fusion of Hindi & English) addressed multiple queries and provided real-time information, e.g., weather info