> Contributing to the development of LinkML, a linked open data modeling language

> Development and maintenance of schema in LinkML for DOE/OSTP-funded National Microbiome Data Collaborative (NMDC) project

> Development of ETL pipelines to integrate knowledge/metadata about microbiome projects from various federal data sources as part of the NMDC data repository

> Development of microbiome related knowledge graphs to power ML/prediction models

> Development of bioinformatics pipelines for the detection of synthetic engineering

> Added APIs/functionality to Python backend/package for a novel data ingress ecosystem, which uses schema.org (serialized as JSON-LD) based data-model specification, and JSON schema for validation

> Packaged/bundled Python app backend, making it more modular and distribution-ready

> Added RESTful web service/interface to expose core module functionalities using Python/Flask, Swagger/openAPI and connexion

> Designed and developed (and benchmarked) algorithm to allow easy modification of nodes/edges on a networkx graph created from JSON-LD

> Integrated Python/Flask backend with frontend powered by R Shiny

> Added unit tests and integration tests using Python pytest

> Added CI/CD workflow to automate processes like linting of code and automatically running the test suite

Developing clinical informatics pipelines and downstream processing automation tools, to help clinical research scientists and coordinators, supporting their work with the NIH-funded All of Us research project.

> Wrote script to automatically update health/patient records with enrollment status in REDCap DB managed by CTSI, from All of Us (AoU) research project's HealthPro system, using necessary APIs.

> Developed automation scripts (using PyCap library) and native REDCap plugins (using "REDCap External Modules" framework) to streamline automatic generation of reports from REDCap based on externally specified (googlesheet) branching logic.

> Migration of scripts from SAS to Python.

> Worked in the Powertrain Mechatronics department—taking care of all (Software in Loop) SiL testing activities like MISRA analysis and MATLAB/Polyspace analysis. Wrote scripts that generate test cases for ECU-specific unit testing modules/functions in the powertrain controller softwares. In addition, I also worked as a web developer, building web-based tools particularly using PHP (frameworks) to automate many processes in the unit testing workflow.

> Worked as a PostgreSQL/PostGIS database developer to write PL/PgSQL, Python (for QGIS as well) scripts to analyze this map data. This map data is used by the Integrated Predictive Powertrain Controls ECU for improving fuel efficiency and optimizing gearshift strategy in Daimler Trucks and Buses.

> Laid out the infrastructure and wrote the software to store and process NOx (emission) information parsed from text dumps of MATLAB/Simulink models, to be used by the technical compliance team (tCMS) in Daimler AG, Germany. Backend was built using MongoDB and Java (for the sake of the PoC).

> Worked with the Protein Data Bank (in Europe, PDBe) team to develop a querying language—MolQL. The ultimate goal of which was the meaningful sharing of annotated structural biology data in the form of 3D views of structures is critical in exploiting data in the PDB. My task was to allow MolQL to parse generic PyMol/JSmol/VMD scripts and manipulate the 3D-model rendered in LiteMol accordingly.

> The basic workflow I employed was — write a set of functions that could interpret PyMol/JSMol scripts, develop a parser that could translate its abstract syntax tree (AST) into a MolQL expression tree, which could ultimately be interpreted by different viewers. Transpilers were written in TypeScript, and visualization tool used a combination of React.js + WebGL.