Ankit Khettry

Industry certified with an experience of over nine years, I am shouldering the responsibilities of a Senior Software Engineer at Flipkart. My interests are in helping businesses harness the value of data using core computer science skills and expertise across the data and ML side of things.

HDP Certified Java Developer - Hortonworks.

Link: http://bcert.me/pdfwrnof

• Data Side of Things: Hadoop Stack, Apache Spark (Batch and Streaming), Kafka, SQL, etc
• Programming languages: Python, Scala, Java
• Cloud: AWS, GCP
• Container Management: Docker, Kubernetes
• Monitoring and Telemetry: statsd, prometheus, grafana, etc.
• VCS and Project Management: Github, BitBucket, CI/CD with Jenkins
• IDEs: IntelliJ IDEA, Pycharm
• Testing APIs: Junit, ScalaTest

Machine Learning Platform, Flipkart

Machine Learning platform components like Hosted Notebooks, Model Training, Model Deployment, Observability and Active Learning.

Decision Assistant, Flipkart

A chat platform meant to help a customer across their end-to-end online purchase journey.

User Purchase Propensity Model, MakeMyTrip

An ML model that predicts a customer's purchase propensity.

Feature Store, MakeMyTrip

A platform that book-keeps, computes and serves useful, measurable pieces of aggregated data useful for analytical and data science needs. Users can configure 'features' using their own compute logic (SQL based or code based), datasets, refresh frequency, and a large number of optional configurable attributes.

Ingestion Framework, Rivigo

A framework used to ingest data in batches from heterogeneous data sources (relational DBs, document DB, etc) into the a datalake. Integrated with AWS Glue for book-keeping. Integrated with AWS Athena for querying.

Clickstream Ingestion and Reports, Myntra

Several batch and realtime applications for ingesting huge volumes of clickstream logs and processing aggregates, which are further used downstream for analyses and dashboards.

Basel 3 Intraday Liquidity (B3IDL), The Data Team

A job consume, process and prepare data for downstream reporting. The ETL outputs are stored into HDFS - to be queried using Hive and for further downstream reporting.

Real time XML Ingestion, The Data Team

A framework meant to consume XML messages from a realtime source (Kafka), validate and consume them based on business rules, and store the results downstream into a noSQL DB (Hbase).

Batch Ingestion Framework, The Data Team

Highly configurable ingestion platform meant to help ingest large volumes of data arriving from heterogeneous sources. Features include validation, in-flight transformation based on configurable business rules, custom data retention periods, etc.