We are actively looking for C++, Python and Java developers to help us extend the MADARA and GAMS middleware for distributed artificial intelligence in robotics. These contractors will be able to join daily standups and scrums with our team members, ask questions and contribute to an active research, development and engineering team that is tasked with building robotics systems that must be reliable, mission-critical, and high performance in outdoor environments.
In general, we are only interested in people who have years of experience in developing well-documented, scalable and high performance code in C++, Java and Python and can work well in a team environment. Read on for details about the simple tests we have created to see if you will be able to quickly integrate into our team and make impact.
Our main need is in C++ developers who can dive into well-documented code and wikis, extract instructions, digest doxygen API information quickly, and get things done. The following is a task that we would like to see done at a minimum before we would consider contracting (est 1-4 hours):
- Install the MADARA middleware
- Create a simple MADARA application that communicates between 2+ nodes using a UDP or ZMQ TCP transport protocol
- Create a github that we can access to look at your code and documentation
- Use MADARA containers in the madara::knowledge::containers namespace to access and mutate knowledge in the KnowledgeBase
- Use KnowledgeBase::send_modifieds to send data
- Be creative and feel free to explore the features
- Save knowledge to a file as either karl, JSON, or binary
- Do NOT use VREP as this may not be working currently and we want you to focus on a MADARA application, not GAMS
- BONUS: Create an on-receive filter that reacts to knowledge changes over the network
- Share the URL to your github and be sure to document your work (especially any installation process we would need to know to run your code). We appreciate this being in a README.md in the root of your github project. Doxygen commenting is especially appreciated.
- Feel free to share/document any issues you have using MADARA or GAMS and what needs to be improved
Pages of Interest: Getting Started | Installation | Containers | Knowledge | Networking | C++ API Documentation | Style and Coding Guide
Our main need in Java developers are people who can work with Android and create tools, interfaces, and extend our JNI port. The following is a task that we would like to see done at a minimum before we would consider contracting (est 1-4 hours):
- Install the MADARA middleware with Java enabled (Android if you want to create an APK)
- Create a sample Java application or Android application that uses MADARA
- Create a github that we can access to look at your code and documentation
- Use MADARA containers in the ai.madara.knowledge.containers package to access and mutate knowledge in the KnowledgeBase
- Use KnowledgeBase.sendModifieds to send data across the network (Multicast or ZMQ)
- Be creative and feel free to explore the features
- Save knowledge to a file as either karl, JSON, or binary
- BONUS: Create an on-receive filter that reacts to knowledge changes over the network
- Share the URL to your github and be sure to document your work (especially any installation process we would need to know to run your code). We appreciate this being in a README.md in the root of your github project. Javadocs are always appreciated.
- Feel free to share/document any issues you have using the Java port and what needs to be improved
Pages of Interest: Getting Started | Installation | Containers | Knowledge | Networking | Java API Documentation
Our main need in Python developers are people who can work with Boost.Python or create tools, interfaces, and extend our Python port and create new, useful features for AI developers and analytics engineers to take advantage of. The following is a task that we would like to see done at a minimum before we would consider contracting (est 1-4 hours):
- Install the MADARA middleware with Python enabled
- Create a sample Java application or Android application that uses MADARA
- Create a github that we can access to look at your code and documentation
- Use MADARA containers in the
madara.knowledge.containersmodule to access and mutate knowledge in the KnowledgeBase - Use KnowledgeBase.sendModifieds to send data across the network (Multicast or ZMQ)
- Be creative and feel free to explore the features
- Save knowledge to a file as either karl, JSON, or binary
- Share the URL to your github and be sure to document your work (especially any installation process we would need to know to run your code). We appreciate this being in a README.md in the root of your github project.
- Feel free to share/document any issues you have using the Python port and what needs to be improved
Pages of Interest: Getting Started | Installation | Containers | Knowledge | Networking |Python API Documentation (help ('madara'))
We are very interested in finding engineers with a good grasp on machine learning who can help us develop tools for autonomy engineers and robotics researchers. The following is a description of tasks that we would like to see done to establish your ability to pursue tasks in classification, regression, clustering or rule extraction (est 2-12 hours, depending on estimated time of running ML algorithms and the size of the data set you want to work on):
- Install MLPACK or TensorFlow
- Install the MADARA middleware
- Download one of the image data sets from DeepLearning.net or a similar site
- Apply classification or rule extraction to the image-based data set.
- Be creative in what you are trying to extract or classify.
- You will discuss your implementation with interviewers. Impress us! If it can be applied to distributed AI in robots, even better!
- Use MADARA to send the classification or rule extraction data
- Create two custom MADARA applications
- One that calls the machine learning functions in MLPACK or TensorFlow and saves useful classifiers or rules into MADARA variables
- One that looks for this machine learning data and displays it or analyzes it in some way
- Feel free to use UDP Multicast or ZMQ transports in MADARA (see Pages of Interest for more info)
- Use containers where possible, such as NativeDoubleVector, DoubleVector or similar data structures on the Containers wiki page
- Use KnowledgeBase::send_modifieds to send data
- The MADARA implementation can be optional if time is limited. You will need to be able to integrate quickly with MADARA though, as it is a primary middleware that we are working on
- Create two custom MADARA applications
- Upload your code to a new github project and provide us with the link
- Feel free to email me with questions on problem statement, installation, or configuration
Pages of Interest: Getting Started | Installation | Containers | Knowledge | Networking | C++ API Documentation | Style and Coding Guide