Avik Basu Email & Phone Number

Sunnyvale, CA, US

Mountain View, California, US

- Leading and developing open source AIOps platform to perform progressive delivery and automated rollbacks- Develop models for multivariate time series anomaly detection

Mountain View, California, US

• Create and build Deep Learning models on massive Operational Time Series data
• Detect anomalies quickly and efficiently to reduce the MTTD during an outage
• Improve developer experience by performing causal analysis on operational incidents

Santa Clara, California, US

• Enhance and upgrade the existing autonomous information block module to incorporate a reinforcement learning based resolution strategy, where each block tries to adapt to each franchise’s needs
• Improved the inference ability of the Order and Order Product information blocks using a softmax regression model to better narrow down to the right order and product given the limited information provided by the shopper

Santa Clara, California, US

• Designed, captured and optimized state transition graphs to analyze shopper’s conversational flow paths in order to improve dialog management and decision making
• Designed and developed product knowledge graphs (in neo4j) comprising of each franchise’s product catalog and features
• Spearheaded and implemented the consultative buying platform for Linc, enabling shoppers to find and select their right choice of products
• Headed and implemented an autonomous information block module consisting of several blocks, each of which having the ability to understand, infer and survey information

Seattle, WA, US

• Created and designed APIs for seller services payment platform using micro-services architecture on the AWS stack
• Enhance the existing anomaly detection system to identify Fraudulent third party sellers
• Selected and volunteered as a Teaching Assistant in the Amazon's internal Machine Learning University

Chicago, Illinois, US

• Processed and enhanced raw 3D vehicle models and their annotations data in order to train and validate them.
• Created methods and visualizations for vehicle landmark alignment and damage detection analysis.
• Modified QA methods for correctly detecting errors and inconsistencies in the machine learning workflow.
• Optimized the way the trained data was stored in the system using an HDF5 approach.
• Enhanced the Machine Learning pipeline to make the learning system more robust.

Los Angeles, CA, US

• Developed a system to detect trends in economics of innovation over time using Topic Modelling analysis and Machine Learning.
• Collected and analyzed massive datasets (using USC High-Performance Computing Cluster) of academic research documents based on their content in order to reveal changes in their underlying structures over time

• Created scripts to unify different types of data from Kinect to a single Point Cloud
• Researched and developed methods for transforming point cloud orientation of complex 3D objects
• Implemented and added custom PCL methods and functions for greater optimization and efficiency

San Francisco, California, US

• Created robust machine learning models for tagging and classifying micro-level images, involving advanced feature engineering and computer vision techniques.
• Developed a data architecture pipeline on AWS encapsulating the core learning algorithms.
• Set up a REST-ful API with Oauth2 authentication, deployed on AWS Elastic Beanstalk server with RDS MySQL backend and S3 storage

Mumbai, Maharashtra, IN

• Enhanced a Rule-Based Machine Translation system that was aimed at translating law documents from English to Hindi with high accuracy.
• Worked on statistical phrase based post-editing, parts of speech tagging, post-edit capture, and data visualization.

Master Of Science (M.S.), Computer Science

Bachelor Of Technology (B.Tech.), Computer Science And Engineering

Higher Secondary School, Science