Danny Depe

Emphasis on neural nets, such as CNNs, RNNs, and GANs

Emphasis on biomedical engineering

• Achieved 40% reduction in test times and 20% increase in first-pass yields by developing near real-time failure Pareto charts within an automated triage system using SQL queries and Tableau calculations
• Eliminated database-related factory downtime to 0 by designing a robust data pipeline with Kafka for streaming manufacturing test data to Snowflake and AWS S3
• Implemented a streamlined manufacturing firmware and test software QA station, catching 95% of FW test escapes before they hit the production floor
• Realized over 30% in cost savings by unifying the engineering validation test ecosystem with a low-cost test setup mirroring the production line test software environment

• Overhauled CI/CD pipeline by migrating from Docker-in-Docker to Jenkins, Buildkite, and Docker Swarm, reducing flaky LiDAR QA regression test failures by 55%.
• Designed and implemented InfluxDB/Grafana dashboards for FPGA build metrics and QA regression test results, enabling 30% faster root cause analysis in camera and LiDAR CI pipelines.
• Programmed Python-based tooling infrastructure for automated AsciiDoc test report generation, simulation vs firmware camera pixel diff visualizations and a JTAG flash tool allowing remote recovery of LiDAR setups

• Implemented functional and Hardware-in-the-Loop (HIL) tests, exercising various communication protocols such as, CAN, ethernet, GPIO, and RS232, on control boards for self-driving cars
• Developed a standardized library of Python drivers for external instruments and proprietary DUTs, improving accessibility and enabling validation and verification for 10+ unique devices
• Built a web app GUI with Flask, Node.js, and Express.js for test operators, validated using Selenium

• Migrated 20kW satellite on-orbit power budget tool from Excel to Python, eliminating 80% of data loss incidents and reducing processing time by 45% for the Spacecraft Contract proposal team
• Characterized and validated a CAN bus with 40+ devices enabling the first adoption of CAN communication on the company’s satellites, saving hundreds of thousands in spacecraft launch costs

I am 100% comfortable coding in Python from scratch. I am comfortable reading C++ and modifying existing code.

• Python
• Pytorch
• SQL
• C++
• Pytest
• Tableau