Anoushkrit Goel

Besides working on my startup, Tensr.AI, I work on real world problems which have the potential to make impact.

I work on problems in Deep Learning, Deep Reinforcement Learning, Machine Learning, Design, Bio-Medical Sciences, and Entrepreneurship. Having started an startup, I have worked on the problems which the startups face in early stage and still I work continually to launch innovative product in the wearable tech domain.

Currently, I am working in the field of Neuroimaging, where our main focus is to construct fiber streamlines and segment fiber bundles, further it expands to denoising the dmri signal and reconstructing the dmri data.

I graduated from Shiv Nadar University with B. Tech in Electronics and Communication Engineering. In my free time, I like to engage in various physical activities.

💻 Professional Experience

(2019 - Present)


Data Scientist

(May 25, 2021 - Jun 1, 2022)

Solved the problem statement of "whether the user will deposit in 5 minutes", now this had 3 objectives , determinstic time dependent prediction of the user behaviour which is based on how the user is interacting with the platform. Interaction further involves decision making while tapping/swiping on the platform. This interaction generates a clickstream containing a lot relevant data where the

Deployed a model which can predict the

IIT Mandi

Deep Learning Research Fellow

(Oct 2021 - Present)

Working on Brain Tract Segmentation at MANAS Lab, IIT Mandi under the guidance of Dr. Aditya Nigam and with Ranjeet Ranjhan Jha. Implemented a combination of Deep Learning Techniques to achieve SOTA accuracy in classifying the Brain Fibers into right classes. Referred multiple research papers for the same.

  • Also, peer-reviewed papers for different journals, and conferences (MICCAI, MIA, ISBI, ICCV)


Github Repos: tract-segmentation, parcellation, unnerve

Absolutdata Analytics

Analyst

(Mar 2021 - May 2021)

Worked on Forecasting Sales Data for FedEx (Client) considering the impact of COVID) and landed FedEx as a long-term client with a 20 million dollar upcoming projects.

  • Implemented various Forecasting like SARIMAX, ARIMA, Prophet.

  • Incorporated various exogenous variables which might have an impact on the sales and have correlation with the abnormality added due to COVID.


IIT Kanpur

NIDHI-EIR Fellow , SIIC IIT Kanpur

(Feb 2021 - Feb 2022)

Worked on scaling up the product Teresa, by Tensr.AI.

Secured NIDHI-EIR Fellowship, for the year to create the hardware prototype for my startup (Tensr.AI), product Teresa which is a Healthcare AI Assistant.

Created the functional prototype of Teresa Band, which enables to capture multiple cardiac modalities for holistic heart health monitoring.

IIT Madras

Pre-incubation, IIT Madras (Jan 2021 - Jan 2022)

  • Finished Customer Discovery Program by Gopalakrishnan Deshpande Foundation (GDC) in which we interviewed more than 75 people about the problems they face on a day to day basis in terms of their health.


  • Mentored by Harish Natarajan, on business, medical & fitness devices.

IIT Mandi

Incubatee, IIT Mandi (Mar 2020 - Feb 2021)

During the Himalayan Startup Trek 2019, our team was selected for incubation at IIT Mandi. The incubation program focuses on making selected startups work towards creating their prototypes.

  • Worked on building the hardware under the guidance of Dr. Shubhajit Roy Chowdhury, also researched on SCG data.

Tensr.AI

Founder and CEO, Tensr.AI (Feb 2020 - Present)

"Increasing Human Potential with Assisted AI"

Developing the first product of Tensr.AI, which is a Healthcare AI Assistant, Teresa. Teresa allows users to triage their Cardiovascular and Respiratory Ailments with the help of the device and services provided by us. The product uses Deep Learning, Digital Signal Processing and other techniques over the bodily signals received by our device.


xtLytics LLC

&

NextGen Invent Corporation

Data Scientist, xtLytics LLC (Jan 2019 - Jan 2020)

Breast Cancer Image Classification and Cancer Grading

Developed an edge-based AI app for the Low Middle Income Countries (LMICs) , primarily Guadalaraja, Mexico, to Triage Breast Cancer using Ultrasound Images with the help of Deep Learning Techniques (CNN) on a data set of around 6500 Images. (from scratch to deployment) with Dr. Susan Love Foundation . Upon this partnership, the team of radiologists and primary health workers helped us annotate, segment and understand the data which led to the creation of an cross-platform mobile application small enough to reside on each primary health workers' phone. This application could detect palpable lumps, and cancerous lesions with the help of a camera image of the scan

Worked with Dr. Susan Love , Dr. Wendie A Berg(MD, phD, FACR, FSBI), Upal Roy, Shobhit Singhal

OCR Recognition

State-of-the-art Invoice OCR using GCP (Google Cloud Platform) which extracts Tables and further transforms the data inside those tables to desired account sheets by the Clients (50 Gas Station in US).

3D Augmented Reality Simulation

For easing the life of new surgeons

📚Education

Indian Institue of Technology, Mandi (Feb 2022 - Present)

Masters by Research
  • Courses

    • Advanced Topics in Deep Learning (CS672) by Dr. Aditya Nigam Feb 2022 - June 2022

Included Transformers, Deep Reinforcement Learning, VAE (Variational Autoencoders), GANs (General Adversarial Networks), GCN (Graph Convolutional Networks)


Shiv Nadar University (Jul 2015 - May 2019)

BTech
  • Major: Electronics and Communication Engineering

  • Minor: Design (Introduction to Ergonomics, Intrduction to Product Design, Color in Design)

  • Other courses in Computer Science and Machine Learning. For the list of courses taken, please visit here

  • Founder VEIG

  • Founder Dauntless


Data Scientist (Jun 2021 - Jun 2022)

Created a Sequential Decision Making Framework which can predict the Deposit Behaviour of a sign up user. We named this as Vision and this allows us to manuever resources towards a likely-depositing user.

  • This framework allows to know future steps in advance, the target event "click" can be any event and the model gives the probability score of occurrence of the same.