Upasana Dhar

1. Built Python scripts to transform binary .silo files into structured HDF5 format, facilitating data access independent of VisIt software. 2. Developed ML classification pipelines on 7,500+ simulation runs, including data cleaning, EDA, class-imbalance handling and classification; researched a neural network to extrapolate simulation outcomes beyond the training boundary.

Conducted requirements gathering and problem definition for ethical AI adoption, synthesising market and technical research into a PoC presented to Deutsche Telekom stakeholders. Defined scope and worked in a cross-functional team across technical and business workstreams.

1. Designed and delivered a PoC to get Chocolatey Nugets stored in JFrog to a front end for 100+ associates to easily install and manage, reducing software provisioning time. 2. Extended an AUTOSAR visualisation tool (Qt/QML, C++, cmake) to parse ARXML parameter descriptions and ECU configuration data, enabling engineers without DaVinci licences to access and inspect critical automotive data.

1. Developed a POC on how to get Chocolatey Nugets stored in JFrog to a front end from where they could be easily installed and managed by 100+ company associates. 2. Designed and developed enhancements for an application using Qt/Qml and C++, that allows associates without a DaVinci license to view ECU configurations.

Full-stack developer in telecom: built a customer-facing GUI (Angular, Node.js, JavaScript) processing 200+ daily orders for a major EU provider. Sole developer on an SNMP proxy tunnel (Python, Linux) handling 150+ devices. Automated a legacy ordering process with UiPath, increasing throughput by 25%. Developed ETL pipelines (Oracle PL/SQL) for a UK telecom client. Mentored new team members and led knowledge transfer sessions. Agile/Scrum. PSM I certified.