top of page

Atharv A. Bagde

Computer Vision | Natural Language Processing | Machine Learning | Network Science

Vision

 

"Where the mind is without fear and the head is held high, where knowledge is free, where the world has not been broken up into fragments by narrow domestic walls


Where words come out from the depth of truth, where tireless striving stretches its arms towards perfection, where the clear stream of reason has not lost its way into the dreary desert sand of dead habit 

Where the mind is led forward by thee, into ever-widening thought and action, into that heaven of freedom, my Father, let my country awake."

~Rabindranath Tagore

Nothing conveys it better than my favorite poem since my childhood

The Journey

My skills across lifecycle of data from its ingestion, analysis, modeling and beyond....

Data  Engineering

Ingestion of data via pipelines marks its inception into the world of Data Science. Here it is securely stored in various databases and warehouses. Pipelines are setup to process and transfer it as necessary

My Skills:

Extensive experience in designing solutions on GCP and AWS databases like RDS, BigQuery, CloudSql, MongoDB and Firestore.

Developed scalable cloud-based ETL pipelines for various applications like streaming-data processing (Apache Kafka, AWS Kinesis, GCP Dataflow) and big data processing (GCP Dataproc and AWS EMR)

Researched on various dataset versioning tools like DVC, Pachyderm and integrated them into ETL and ML pipelines.

Adept with de-identification, ingestion, storage and integration of medical data formats like FHIR, HL7 and DICOM using GCP Cloud Healthcare API.

Built a pipeline that identified false positives returned from GCP Cloud Healthcare API that detects images with PHI information on it.

Data Analytics

Next step is to visualize and analyze the ingested data.

Data is cleaned, transformed and made ready with newly calculated features for modeling

My Skills:

Integrated visualization tools like Tableau, PowerBI and Looker to cloud databases on GCP, AWS and Azure.

Excelled at graduate-level statistical inference course covering topics like regression, ANOVA and hypothesis testing.

In depth knowledge of probability distributions and Bayesian probability concepts

Extensive experience in analyzing and visualizing medical image formats of DICOM and NIFTI using open source softwares like 3D slicer.

Adept with feature engineering techniques for both temporal and relational data-sets.

Machine Learning

Pivotal step! The cleaned and processed data is modeled using traditional ML algorithms or Deep Learning methods.

My Skills:

Well versed with​ building 2D and 3D Convolutional Neural Networks for object classification, detection and segmentation using U-Net, YOLO and Faster-RCNN architectures

Designed and developed solutions for sentiment analysis, machine translation and sentence completion using LSTMs and BERT

Developed a parametrized 3D GAN for generation of realistic CT images based on user input

 

Proficient in ​deep learning frameworks of Tensorflow and PyTorch with 3+ years of experience. Currently exploring the JAX framework.

Technical experience and elementary research in novel techniques like Federated Learning and Self-supervised Learning.

Highly skilled in employing traditional ML algorithms of SVM, XGBoost, Random Forest, DBSCAN, KNN and K-means clustering

Researched and experimented with loss functions like CL-Dice, Betti number and Hausdorff distance for 3D thin-object segmentation as part of a NIH funded project called HubMap

Experience in building and deploying Document-OCR and object detection models using GCP AutoML.

ML OPS

Components...Assemble! After all the components of the project are built, they are fixed together in a pipeline to ensure a smooth and scalable execution.

My Skills:

Integrated tools that enable experiment tracking, metric visualization and dataset versioning into a deep learning training pipeline

Developed a Python wrapper to train deep learning models using multi-node distributed learning with multi-gpu or TPU accelerators.

Led the project to design, build and deploy an end-to-end MLOPs pipeline using Kubeflow-based Vertex-AI pipeline on GCP

Co-authored a  guide for the best practices and tools used for design, development and deployment of a deep learning based SAMD (software as a medical device)

Well versed with model compression and inference optimization (Tensor RT, ONNX). Currently experimenting with GPU and deep-learning focused languages like Titron and Julia.

pipeline-stock_edited.png

Management and Sales

Data's journey culminates at this stage where an application/system modeled on it is pitched as a solution to a client's problem. To ensure viable and efficient development of such applications, project management is of utmost importance

My Skills:

Acted as the technical consultant for multiple proposals of the Go-to-Market (GTM) team across varied domains like self-supervised learning, predictive maintenance, and image indexing.

Converted 8 out of 10 sales pitched and effectively generated revenue of $1.5 million+ for the company

Drafted Statement of Works for multiple proof of concepts and development projects.

Well versed with different project management methodologies like agile, kanban and waterfall using JIRA and Azure Boards

Prepared detailed proposals with project timelines, team requirements, deliverables list, cost estimations and other dependencies in response to numerous Request for Proposals.

Integrated Azure Boards with development pipeline to automate ticket-updating, project documentation and  metrics visualization.

Professional Certifications

GCP-ACE.png
576137f8-pmle_edited.png
IIT_Madras_Logo.svg.png

Testimonials

1578651292719.jpeg

Rahul Sreenivasan

Engagement Manager -- Quantiphi Analytics
 

 

I have had the pleasure of knowing Atharv for more than 2 years now.  He has worked in a client facing role as a lead on several Machine Learning projects ranging from Natural Language Processing (NLP) to Computer Vision (CV) and Optical Character Recognition (OCR)

He has exceptional interpersonal skills, a strong work ethic and a positive attitude towards work. He is diligent in that he ensures that the client requirements are understood well not just by him, but by everyone in the team and strives to deliver above and beyond all expectations. To top it all, he is a great person to have on the team. Atharv is an amazing person to work with. Highly recommended!

1666706759267.jpeg

Pratik Bijam

Principal Technical Architect - Enterprise AI/ML

Quantiphi Analytics
 

 

Atharv is one of the best machine learning engineer in our team. He has played supportive role during the time of growth for the team which is highly appreciated by the leadership. In a project full of talent, Atharv stood out because of his exceptional technical and behavioral skills. Atharv always brings his point of view to the table which makes solutions interesting, resulting in high client satisfaction. I wish Atharv all the best for his career..!!

bottom of page