Anindya Sudhir
Machine Learning Engineer/Data Scientist
Pune
Summary
Machine Learning Engineer/Data Scientist skilled in Machine Learning & AI with11+ years of experience in IT industry as Software Developer and Data Scientist/Machine Learning Developer. Have excellent real world problem solving ability using Machine learning/Deep Learning Experience in Identifying use cases, selecting tech-stack and developing solutions using Data Science. Extensive past experience in Software Development and Agile way of working.
Overview
11
11
years of professional experience
7
7
years of post-secondary education
3
3
Certifications
Work History
Machine Learning Engineer
Vodafone
Pune
11.2021 - Current
- Working in Advanced Analytics team for Fraud Detection project for Vodafone Germany.
- Leveraged artificial intelligence and machine learning algorithms like Xgboost, H2o AutoML, and Autoencoders written using Pytorch for Fraud detection.
- Identified new problem areas and researched technical details to build state-of-the-art ML model.
- Currently building the ML pipeline to implement incremental learning to tackle the drift in data.
- Worked on the near real-time Fraud detection pipeline written in Scala, Spark, Sparkling water(H2o AutoML), Kafka, and Cassandra for real-time fraud predictions.
- Worked on the fuzzy matching logic to identify returning customers using RedisSearch (for performance).
- Working on building realtime pipeline for fraud detection on event driven microservice architecture using Kafka
- Responsible for building end to end ML pipeline, from data sourcing to feature engineering to ML model training and deployment to predictions.
Machine Learning Engineer
IBM
Pune
04.2018 - 11.2021
- Built Fine Tuned BERT (Base uncased), ALBERT, DistilBERT etc models using Huggingface and SimpleTransformers libraries for sentence classification task and performance of all models were compared. Data available was highly unbalanced so, several class balancing techniques were tried like Undersampling, Oversampling and Class weights to get better F1-score. Model training was done in IBM cloud pak for data for GPU access.
- Developed RASA based Chatbot for helping support team. End users interact with bot to fetch relevant information from asset management applications and log ticket on ticketing platform in case user wants that. Saved ~20 hours/week of support team.
- Identified and analyzed Data Quality issues in data pushed to IBM's Data Lake CEDP and made entity recognition system using RASA framework, API calling and returning correct entity suggestion as response back to end user.
- Mood Marbles Sentiment Analysis
Software Developer
IBM
Pune
04.2017 - 03.2018
- Developed robust Java code for IBM asset management applications using latest technology stack which includes Spring Core, Spring Security, Spring Data JPA, MongoDB, Postgresql, Docker, Jenkins for CI/CD etc.
- Responsible for code review of peers
- Deployment of application in end of sprint using Jenkins
- Increased Performance of application by Query optimization, Caching, non blocking APIs and moving from HTTP1.1 to HTTP2
Software Developer
Cognizant
Pune
09.2013 - 03.2017
- Developed module which gets data from Oracle database and publish it in form of tree in application’s frontend
- Ensured that tree loading time is less than 1 second by implementing Lazy Loading in Tree
- Conversion of Parent-Child relationship data to Dynamic JSON by using Depth First Algorithm
- Query optimization: Assessed execution time of queries and Re-wrote (optimized) queries to reduce execution time
- Developed module which gets data from Oracle database and publish it in form of tree in application’s frontend
- Ensured that tree loading time is less than 1 second by implementing Lazy Loading in Tree
- Conversion of Parent-Child relationship data to Dynamic JSON by using Depth First Algorithm
- Query optimization: Assessed execution time of queries and Re-wrote (optimized) queries to reduce execution time
- Collaborated with project managers to select ambitious, but realistic coding milestones on pre-release software project development.
- Revised, modularized and updated old code bases to modern development standards, reducing operating costs and improving functionality.
Education
Master of Science - Machine Learning & AI
Liverpool John Moores University
Liverpool , England 04.2020 - 12.2021
Postgraduate Diploma - Machine Learning & AI
International Institute of Information Technology
Bangalore 04.2020 - 04.2021
Bachelor of Engineering - Information Technology
Birla Institute of Technology, Mesra
Ranchi 04.2009 - 04.2013
Skills
Machine learning/AI
Python
Scala
Cassandra
Kafka
Redis
Redisearch
Keras
Sql
Tensorflow
Pytorch
Cassandra
Spark
Kubernetes
Docker
Jenkins
Certification
Machine Learning, by Stanford University and offered through Coursera
Timeline
Machine Learning Engineer
Vodafone
11.2021 - Current
IBM Cloud Core
07-2021
Master of Science - Machine Learning & AI
Liverpool John Moores University
04.2020 - 12.2021
Postgraduate Diploma - Machine Learning & AI
International Institute of Information Technology
04.2020 - 04.2021
Machine Learning, by Stanford University and offered through Coursera
08-2019
Oracle Certified Associate, Java SE 8 Programmer
12-2018
Machine Learning Engineer
IBM
04.2018 - 11.2021
Software Developer
IBM
04.2017 - 03.2018
Software Developer
Cognizant
09.2013 - 03.2017
Bachelor of Engineering - Information Technology
Birla Institute of Technology, Mesra
04.2009 - 04.2013