Aug 2021 – Feb 2022 · 6 mos

• Lead the Data Engineering and Machine Learning Team to build projects like HS Code classification, Merging of duplicate companies (Entity Resolution), TEU and shipment value estimation, Semantic similarity search for product keywords, Similar buyers and suppliers, Trade Risk Prediction using Denied Parties List, Country-wise data pipeline creation along with derived attributes computation using the raw fields.
• Technologies used: Python, Scikit, PyTorch, Huggingface, Apache Airflow, Elasticsearch, MongoDB, PostgreSQL, Neo4j, Grafana

Oct 2020 – Aug 2021 · 10 mos

• Building Data and ML Platform from scratch @Trademo

Sep 2019 – Oct 2020 · 1 yr 1 mo

• Part of the ML team under Consumer Technology group. Projects:
• 1) Tagging and Sentiment Analysis of OYO reviews: Auto-tag user reviews related to Room, Washroom, AC etc. along with the sentiment using RoBERTa pre-trained model and XGBoost in Python using Scikit, NLTK etc. which reduced the dependency of manual classification.
• 2) Auto-Classification of OYO Images: Classification of OYO images with a Deep Neural Network using transfer learning (ResNext) with PyTorch which reduced the dependency of manual classification, for 85% of the hotels.
• 3) Hotel Content Scoring: Compute scores of OYO Hotels on the basis of their content to facilitate auditing and boosting good quality content hotels in the Search ranking algorithm using PySpark and Scikit which helps the hotels to improve and their content and get a boost in their ranking.
• 4) Cancellation Prediction Model: Data sourcing, Pipelining, Feature Selection, Pre-processing, Model creation, Deployment and Monitoring using XGBoostClassifier in Python which led to jump in URNs and decrease in cancellations in Indonesia and Malaysia markets.

Nov 2017 – Sep 2019 · 1 yr 10 mos

• Part of the Airtel Customer 360 vertical facilitating user data in real time. Projects:
• 1) Airtel Logging and Monitoring Framework: Standardisation of Log creation, storage, parsing, information extraction and dashboard creation using FluentD, Telegraf, InfluxDB, Elasticsearch and Grafana. This framework became the logging standard for all projects across Digital Intelligence Team.
• 2) Customer 360: Creation of user data snapshot pipeline for 600 million Airtel customers using Apache Kafka, Aerospike, Elasticsearch and Python. The system was used by 50+ internal services of Airtel to fetch user data attributes.

Jul 2017 – Nov 2017 · 4 mos

• Data Pipelining, storage, batched and real-time processing using Firebase, Apache Airflow, Elasticsearch, RabbitMQ and BigQuery. User personalization and feed optimization.
• 1) Feed Optimization : Personalisation of user feed with overlapping K-Means clustering using Python, BigQuery and Elasticsearch. This improved the average time spent and CTR of the posts shown in the Feed.
• 2) Analytics pipeline : Logging of user events using Firebase, initiating activity based triggers and setting up an analytics pipeline using Airflow and BigQuery. This project formed the basis of data monitoring systems and helped executing well-informed decisions.

Aug 2016 – Jul 2017 · 11 mos

Jul 2015 – Aug 2016 · 1 yr 1 mo · New Delhi Area, India

• Part of the Data Platform and personalisation team.
• 1) Automation of user personalization and recommendation related projects: Trending Now, Frequently Bought Together, Similar Products using Spark, Cassandra and Kafka.
• 2) Customer Experience Engine that sends personalized push notifications to users based on their past activity.