Summary
Overview
Work History
Education
Skills
Timeline
Training And Conference Attended
Training And Conference Attended
Training And Conference Attended
Generic

Kumar Deepankar

SAN DIEGO,CA

Summary

As an AI and ML Engineer lead at Tata Consultancy Services, I have 11 years of experience in building and deploying data, ML and AI solutions for enterprise clients, especially in the natural language processing, machine translation and image classification domain. I have helped client streamline their data pipeline to manage massive data for meaningful business outcomes in domain of search technologies, distributed processing of large data, finding patterns /generating insights from data.

I have a strong background in application development, big data technologies, cloud infrastructure, distributed systems, machine learning, developing / evaluating state of the art neural networks.

Overview

11
11
years of professional experience

Work History

AI/ML lead

Pfizer Inc.
10.2019 - Current
  • Working as technical lead in Artificial Intelligence and data science domain
  • Driven global teams in agile working environment with hybrid onsite-offshore model to achieve operational efficiency
  • Involved in successful architecture design, collaborated with stakeholders to incorporate recommendations into design
  • Architected high-level and low-level systems
  • Assessed viability of emerging technologies to meet business need
  • Built, trained, and evaluated Neural Machine Translation (NMT) models
  • Deployed ML solutions (Translation and Transcription) as services
  • Worked on services qualification that met regulatory needs to enable enterprise wise adoption
  • Streamlined MLOps pipeline for model evaluation against industry standard metrics that enabled crucial decision making
  • Worked to evaluate ML model viability
  • Brought down the training cost of deep learning models by about 25% by refactoring code to utilize multiple GPU set up
  • Developed reverse proxy micro-services REST APIs to integrate authentication and authorization, metric logging, and subscription microservices to eliminate redundant services required across solution deployment
  • Developed API gateway services for multiple back-end services that enabled vendor agnostic goal
  • Deployed data services on Kubernetes cluster to reduce the operating cost of applications
  • Packaged Elasticsearch and Apache Solr in container images
  • Set up micro-services edge stack and cloud services in cost effective manner.

Pfizer Inc.
10.2016 - 09.2019
  • Designed, developed, and analyzed deep learning architectures on clinical images (Rat organs scan) to classify the organs cells as normal or diseased
  • Significantly improved the training time of models by re organizing the code to work on multi–GPU EC2 servers of scale p3.16xlarge (eight NVIDIA Tesla V100)
  • Evaluated and compared popular CNN architectures AlexNet, ResNet, GoogleNet and VGG on clinical images
  • Worked on miscellaneous activities on images
  • Extracted scale invariant features from images
  • Used images filter to extract features
  • Performed low level image processing tasks like edge and blob detection
  • Evaluated machine learning algorithms on the extracted image features for image classification
  • Mapped the clinical images and displayed using Neo4j
  • Build and evaluated models for predictive modelling in RapidMiner
  • Set up Elasticsearch cluster on AWS and provision the cluster in an enterprise set up
  • Performed operational tasks on Elasticsearch cluster not limited to Multitenant support, Authentication and Authorization, back up operation, Disaster recovery set up
  • Designed indices mapping and perform indexing operations on NLP Analyzer of Elasticsearch
  • Design custom NLP Analyzer for index/search specific needs
  • Worked on data pipeline to pick laboratory files originating from electronic notebook and store in AWS cloud storage system
  • Designed index mapping to search laboratory files stored in AWS cloud storage

Senior Developer

AIG
04.2014 - 09.2016
  • Cleaned the adjuster notes text
  • Performed NLP operations on the adjuster notes text to extract entities not limited to DOB, SSN, address, claimant, insurer
  • Indexed the processed text in to Apache Solr
  • Improved search relevancy
  • Worked on call center audio files
  • Converted audio to text and subsequently built machine learning models to classify the phone calls
  • Worked on NLP to process text file generated from audio.

Big Data Developer

TATA Consultancy Services
03.2013 - 03.2014
  • Part of Digital Transformation Center of Excellence team
  • Implemented digital solutions for various clients
  • Presented various Proof of Concept in MapReduce to improve turnaround time of various ETL batch jobs
  • Worked on MapReduce / Apache Hive/ Apache Pig and other components from Hadoop Ecosystem
  • Worked on MapReduce framework for reading and processing files of different format
  • Worked with low level api for custom input format
  • Re-written custom code in MapReduce to improve the query relevancy in Hive and Pig
  • Worked on custom implementation of popular machine learning algorithms viz Naïve Bayes, Decision Tree, Random Forest, KNN and other statistical NLP algorithms
  • Explored statistical NLP algorithm viz Association rules, Frequent pattern mining, opinion mining, and recommendation systems.

Education

Master of Technology in Data Sciences -

IIT, Hyderabad

Bachelor of Engineering in Electronics and Communication Engineering -

SLIET, Longowal

Skills

  • Python
  • Java
  • Go
  • AWS (S3, Redshift, DynamoDB, Snowball, EMR, ECS, EC2, ELB, OpenSearch)
  • Distributed Systems (Elasticsearch, Hadoop Ecosystem, Kubernetes, distributed storages, Apache Cassandra)
  • Docker
  • TensorFlow
  • Keras
  • Machine Learning
  • Deep learning networks (Popular CNN and LLMs)
  • Data Modeling REST API
  • Python microservices framework (Flask and Django)
  • CI/CD (GitHub /Jenkins /Airflow/ Spinnaker)
  • ELK Stack
  • RapidMiner
  • QlikView
  • Tableau
  • Tensorflow
  • Pandas
  • NumPy
  • Higging Face’s Transformer
  • LangChain
  • Insurance
  • Pharmaceutical

Timeline

AI/ML lead

Pfizer Inc.
10.2019 - Current

Pfizer Inc.
10.2016 - 09.2019

Senior Developer

AIG
04.2014 - 09.2016

Big Data Developer

TATA Consultancy Services
03.2013 - 03.2014

Master of Technology in Data Sciences -

IIT, Hyderabad

Bachelor of Engineering in Electronics and Communication Engineering -

SLIET, Longowal

Training And Conference Attended

  • 2013 Computing for Data Analysis, Coursera
  • 2017 Neo4J developer training, Neo4j, New York
  • 2017 Elastic {ON}: Annual Elasticsearch conference, San Francisco
  • 2018 Elasticsearch Engineer -1 training, NewYork
  • 2018 RapidMiner Developer, online
  • 2021 KubeCon + CloudNativeCon North America 2021, Los Angeles
  • 2023 DeepLearning.AI TensorFlow Developer

Training And Conference Attended

  • 2013 Computing for Data Analysis, Coursera
  • 2017 Neo4J developer training, Neo4j, New York
  • 2017 Elastic {ON}: Annual Elasticsearch conference, San Francisco
  • 2018 Elasticsearch Engineer -1 training, NewYork
  • 2018 RapidMiner Developer, online
  • 2021 KubeCon + CloudNativeCon North America 2021, Los Angeles
  • 2023 DeepLearning.AI TensorFlow Developer

Training And Conference Attended

  • 2013 Computing for Data Analysis, Coursera
  • 2017 Neo4J developer training, Neo4j, New York
  • 2017 Elastic {ON}: Annual Elasticsearch conference, San Francisco
  • 2018 Elasticsearch Engineer -1 training, NewYork
  • 2018 RapidMiner Developer, online
  • 2021 KubeCon + CloudNativeCon North America 2021, Los Angeles
  • 2023 DeepLearning.AI TensorFlow Developer
Kumar Deepankar