Jan 2024 – Dec 2024 · 11 mos · Bengaluru South, Karnataka, India · Hybrid

• Used ensemble techniques to Improve profile matching accuracy of medical profiles from 38% to 82%.
• Fine-tuned transformer based model with 94% accuracy to classify job-roles from input job-titles.
• Used embedding techniques to classify a candidate work experience into 27 different job-functions.
• Used ChatGPT for data set curation for job-role and job-function classifier.
• Created data pipelines using Airflow for ingesting and invoking ML classifiers for mapping candidate profiles.
• Tools - Python, Airflow, ChatGPT, MongoDB

Jan 2022 – Jan 2024 · 2 yrs · Remote

• Created a serverless data lake for a unified source of accurate business metrics..
• Used Generative AI techniques to generate engaging short stories and accompanying images for kids.
• Implemented a Subscriber Knowledge Graph to enhance game recommendation and predict churn.
• Scheduled 90% of critical data scripts on Airflow and configured alerts to get success/failure notifications.
• Leading the data team of 5 members to drive data driven decisions in the organization
• Tools - Python, Apache Airflow, AWS Glue, AWS Athena, AWS S3, BigQuery, MySQL, MongoDB, Neo4j, Generative AI, GPT3, LLMs

Feb 2020 – Jun 2021 · 1 yr 4 mos · Greater Delhi Area

• Added Text Classification & Named Entity Resolution modules supporting transformer architectures like BERT, RoBERTa, XLNET, etc to our enterprise software offering arcgis.learn.
• Used embedding technique to find similar images from a corpus of 9.5 million images.
• Trained a NER model and created a Knowledge Graph by extracting named entities from text documents.
• Trained GPT-2 transformer model to generate meaningful text for a given incomplete sentence/paragraph.
• Tools - PyTorch, FastAI, Faiss, TensorFlow, HuggingFace Transformers, spaCy, Neo4j

Jan 2018 – Jan 2020 · 2 yrs · Noida, Uttar Pradesh, India

• Involved in engineering and creation of various Data Science related services and pipelines.
• Created and productionized multiple Knowledge Graphs built on Neo4j using Docker.
• Involved in re-architectures of our data ingestion pipeline so as to make it more fault tolerant.
• Tools - Python, Neo4j, Docker, Java, MongoDB, RabbitMQ, Redis

Jan 2017 – Dec 2017 · 11 mos · Noida, Uttar Pradesh, India

• Working as a data engineer to create a resume to job match engine which recommends jobs based on a user's resume.
• Created an automated pipeline for our resume to job semantic match API (via Doc2Vec approach).
• Created REST APIs to serve enriched and insightful data to internal teams.
• Involved in optimizing and reducing the run time of Data Science APIs.
• Involved in setting up Mongo clusters and RabbitMQ broker system for our test and production
• environments.
• Tools - Python, Java, Gensim, MongoDB, RabbitMQ, Azure Service Bus, Shell Script, Redis, WebServices, Jersey, Twisted framework

Aug 2015 – Dec 2016 · 1 yr 4 mos · Noida, Uttar Pradesh, India

• Written software robots to extract product information from various e-commerce websites in the US market.
• Done continuous improvement in the platform to allow data extraction from new e-commerce websites.
• Carry out various POCs at both infrastructure and platform/code level.
• Tools - Python, Selenium, Amazon Redshift, Amazon EC2, Google Cloud, MongoDB, shell script

Sep 2013 – Aug 2015 · 1 yr 11 mos · Gurugram, Haryana, India

• Worked as a Django developer to build a portal which provides information on various colleges across India.
• Worked on Apache Solr to improve the search quality of the site based on the keywords entered by the user.
• Tools - Python, Django, Solr, MySql, and Memcache

May 2012 – Dec 2012 · 7 mos · Lake Mary, Florida (USA)

• Wrote Python modules to call these web services from respective URLs.
• Creation of webservices in Python & C++ to collect system information related to Rating & billing software.
• Modifying the existing web Services to add new parameters required by the system.