Ankur Chavda

Developed data pipelines that fed data into machine learning models for outlier detection using PySpark, SQL, Python, Shell Scripting. Pipelines to clean and supply news data, Reddit posts, comments and activities in other such forums to the machine learning models using Python, Couchbase, NoSQL, Flask, Docker. REST APIs for providing sentiment trend data of various companies on a dashboard used by senior leadership. Reputational Risk dashboard that monitors a surge in negative sentiment on Twitter for American Express to address issues proactively. Led the team and set coding standards by developing common packages, git workflow and review processes.

Transformed and processed huge sets of Real-World data to OMOP format for pharmaceutical giants in the US using Spark, Python, SQL, Databricks, AWS. Built pipelines using Airflow to generate cohorts for analysis. Automated and optimized transformation runs on monthly and quarterly data feeds using Databricks, Spark, SQL, Python. Led the migration of a Teradata ecosystem to Spark and AWS starting from design, scripts, automation, conversion of data to parquet format and query compatibility.

Developed an email scheduler for a Danish clothing & accessories giant for targeted email campaigns using Node.js, Express.js and MongoDB.

Enhanced a hedge fund & portfolio management software for a US based financial services company in C#, Angular JS.