Antmicro offers students the opportunity to put their knowledge to practical use by writing their Bachelor’s or Master’s thesis with our company. Interns who participate in Antmicro’s Internship Program and decide to focus on open source technologies in their final university project or thesis are offered various forms of support:
- Access to Antmicro’s latest technologies
- Technical expertise provided by Antmicro’s engineers
- Equipment needed to build a prototype if your project needs it
- Access to oscilloscopes, logic analyzers and RF measuring tools for laboratory prototype design and testing
- Equipment for rapid prototyping of electronic devices including a 3D printer
On top of that, the most talented graduates with the best projects may be awarded a graduation scholarship. To get more information on the types of projects you can work on, take a look at these examples:
SDI-MIPI Bridge
One of Antmicro’s interns created a device that enables to connect industrial and filmmaking cameras with an SDI interface to embedded processing platforms accepting the MIPI CSI-2 standard. The prototype device has a dedicated SDI deserializer combined with dedicated MIPI PHY and featured programmable logic. This solution makes it possible to process the signal on the fly. It also has a variety of applications such as interactive overlay generation, scene analysis, gaining image statistics, Edge AI-enhanced follow focus and many more. The hardware platform was designed in KiCad and is released as open source hardware.
Analysis of execution metrics in TensorFlow Lite machine learning algorithms with the Renode environment
This thesis project focused on analyzing the characteristics of machine learning algorithms and discovering their correlations with software execution metrics such as executed instructions, memory or peripheral accesses and exceptions. The design was based on popular open source tools. TensorFlow and TensorFlow Lite libraries were used for training models and deploying to IoT devices. Renode provided the environment for functional simulation, enabling debugging and testing of unmodified embedded software on a PC while the collected execution data was processed by matplotlib - a python data visualization library.
Synchronization of open source processors in the AMP architecture
Proper task coordination and efficient communication between heterogeneous processors is one of the most important challenges in multiprocessor systems. Although there are many possible solutions to this problem, the available open source processor implementations on FPGA chips were limited. The proposed solution had to be universal, independent of rarely available features and accessible for heterogeneous processors. Research in this area was based on the open-source LiteX framework which allows convenient building of FPGA-based SoCs.