- Pipes: A platform for building peer-to-peer applications. We've been working on this on-and-off for a few years now, while also using it in our own infrastructure as well as for client projects.
- P2P collaboration tool: our first p2p application built on the Pipes platform. Our team has used it exclusively for working together remotely since the summer of 2020, replacing Slack, Skype, Google Drive, Github, and Fogbugz -- and adding new custom development tools.
- CRDT Application Server: An application server framework built using principles borrowed from our Pipes technology, which provides realtime synchronization and support for offline clients. Includes Python, iOS, Android, and Javascript SDKs for building client applications.
- Technical Self-Sufficiency: we've made great strides in our ongoing quest to reduce our dependency on the technical juggernaut. No more AWS, Slack, Github, Dropbox, FogBugz -- we create and manage our own VMs on our own blade server, use Pipes for all of our networking and other internal communications, and our homebuilt collaboration tool for source control, code review, case/milestone tracking, case/commit linking, source/commit browsing, change notifications, task management...well, everything we can dream up wanting it to do.
- Cases: Using the CRDT application server, we've built our own case management system. It takes what's good from the many tools we've built in the past, and throws away the other 95% which was terrible. All of our work is coordinated with this tool -- fully integrated with our P2P Collboration Tool, and it's nice to finally have it under our complete control.
- Logging: Perhaps the most useful developer tool we've ever built.
- R20S: Yup, we're (finally!) building our own operating system. The goal: spend 50% or more of our computer time in R20S, in which every line of every piece of software will have been written ourselves. Partly an exercise in self-reliance, partly an experiment to see what an OS developed without any historical cruft might look like....mostly a whole lot of fun.
- Al: the programming language we have always wanted, and have finally started to use. Al was built from the start to support the design and architecture of large-scale, complex software projects, in a way that no other programming language (that we know of) has ever done.
- Memory dump: a client project to dump the contents of physical memory. Easy, right?
- ClearCrypto: our custom cryptosystem designed for a peer-to-peer internet.
- The biggest matrix: As part of our ClearCrypto project, we're pretty sure we've performed the largest matrix-vector product ever: in over 16 million (224) dimensions. Optimization fun.
- Segment Processor: Our new general-purpose bioinformatic tool intended for analysis of any Illumnia sequencing experiment, which accurately handles adapters, barcodes, linkers, UMI, toggle mutations, and indels.
- SPATS: An open-source bioinformatic tool that performs high-speed sequence alignment and corresponding statistics for analyzing RNA structure.
- SPATS performance optimization: An overview of how we achieved 1000x performance improvement -- while at the same time improving the accuracy analysis -- in the SPATS project.
- Never trim adapters: If you're using one of the many adapter-trimming tools as part of your DNA sequencing pipeline...don't!
- Barcodes: Does your sequencing experiment make use of indexes/barcodes/UMI? This is an overview of the barcode analysis that is part of our Segment Processor tool.
- Cas13b motifs: While assisting with early COVID-19 research, we performed genomic sequence mining for CRISPR-Cas-like proteins.
- Overview of a machine-learning security sandbox we designed and created for a ML/AI startup client.
- Overview of a custom blockchain we designed and created for a ML/AI-on-blockchain startup, including a study demonstrating scalability to millions of users.
- We ran design and development for a proprietary Autonomous Driving (AD) simulator, achieving significant speedup and accuracy improvements.
- Custom spreadsheet-like software: We designed and created a custom commercial real-estate market analysis tool, with a spreadhseet-like UI. Replaces hours of manual work and several overly-complex Excel spreadsheets with a user interface designed specifically for the client's research-intensive activity.