Oct 2021 – Oct 2023 · 2 yrs · Bangalore Urban, Karnataka, India

Apr 2021 – Jun 2021 · 2 mos · Ahmedabad, Gujarat, India

• Built unsupervised algorithm to identify sentiments used guided LDA, NRC sentiment, Loughran McDonald score on post scrapped (using PRAW) from wallstreet bets/reddit on trading
• Identify the words create more engagement, applied topic modelling and effect of tickers on stock prices with 91.4% accuracy for most helpful comment

Jan 2021 – Present · 5 yrs 4 mos · Bengaluru, Karnataka, India

• ListenAI is a conversational-based AI platform that helps to improve agent performance with automated speech recognition and by use of Natural Language Understanding (NLU). We help agents to capture every single interaction, analyze customer interaction, behaviour changes, emotion identification whether being Voice or text or video and provides a simple, scalable solution to deliver a bottom line impact.
• We are a team of IIT Kharagpur graduates with diverse experience in MNC's and very lucky to have our advisors from IIT KGP Alumni working at Mckinsey, Bain and company and Morgan Stanley. We’re passionate about voices, language, technology and results, and can't wait to find moments in conversation that boost the performance for your team.
• If you would like to join us or know more about us, feel free to reach out to me.
• Website: www.listenai.in

Apr 2020 – Jun 2020 · 2 mos · Bengaluru, Karnataka, India

• Worked on Image Processing, OCR techinque to find entities in the land records for the Karnataka state and build a custom NER Model to match entities and NLP algorithms to summarized the document

Dec 2019 – Jan 2020 · 1 mo · Singapore

• Worked on a computer vision project to determine the waiting time in a restaurant by building a object detector and object tracker on a live video.

May 2019 – Jul 2019 · 2 mos · Mumbai Area, India

• Build an unsupervised algorithm to determine the sentiments of audio calls by extracting audio features to determine emotions, used HMM model and neural networks
• Achieved an accuracy of 82.14% to correctly flag the calls on basis of its sentiment and 74% for speaker diarization when tested over live audio calls

Jul 2018 – Jul 2020 · 2 yrs · IIT Kharagpur, India

• Leading a team of 30+ data and consult enthusiasts in campus, worked on real life problem across Finance, Economics, Energy, Innovation and Strategy.
• Being a member of Global Student Think Thank working with 600+ members across 12 countries delegated to Business Strategy network and responsible for making propositions and tracking business solutions globally.

May 2018 – Jul 2018 · 2 mos

• Project 1: Used CNN (Inception architecture) to detect the adult content and achieved an accuracy of 95.16% for 1000 test images
• Project 2: Used CNN (ResNet50 architecture) to detect the type of tumor diagnosis(benign or malignant) with an accuracy of 97.16%
• Used data augmentation to evenly distribute the data and also to increase size of the dataset which resultantly increased the accuracy

May 2018 – Jun 2018 · 1 mo · Udaipur, Rajasthan, India

• • Developed an algorithm to predict predict short-term credit ratings of the companies used default ratios and risk measures like Alfa and beta