Larry Vincent

"We’re going to need a bunch of SCARs, and we’re going to need them as fast as we can make them.” - Dr. Kelly Hammett, director of The Space Rapid Capabilities Office for the United States Space Force. AV will soon be rolling our Badger phased array antennas off the manufacturing floor as part of the $1.4B SCAR program. Read more from Breaking Defense to discover how Badger will transform our nation’s satellite command and control capabilities⇢ https://lnkd.in/gECaSpPz

309

5 Comments

🚨 I see this mistake all too often: You want your robot to move autonomously. You dive straight into setting up Nav2. Then you hit an issue and say: "Nav2 isn’t working." But here’s the thing - Nav2 is a top-layer package. It relies heavily on the correctness of the underlying layers. If those aren’t solid, navigation won’t be either. Before blaming Nav2, make sure these layers are in place and working: * Robot Description: The robot’s physical configuration is accurately defined. * Control: When you send a velocity command, the robot actually moves accordingly. * Localization: The robot can reliably determine its position in the environment. 👉 Get these right first and Nav2 will start to shine. #ROS2

355

16 Comments

DroneShield to expand AI R&D in Virginia for DoD alignment DroneShield has announced it will significantly expand its research and development operations in the United States, more than doubling its U.S. workforce and adding a second location to its headquarters in Virginia. The Australian-listed company said over 30 percent of the new positions will focus on software development, particularly in artificial intelligence. The expansion is intended to accelerate product development cycles and improve support for U.S. customers, according to company officials. #CounterUAS #AI #ArtificialIntelligence #Research #Drones #Defense https://lnkd.in/eGkBKTjU

42

Some real truths as a seasoned engineer using copilot a lot for the last few months: - my productivity has easily gone up 20-30%. The rate at which I’m solving problems has become exponentially faster. No, I’m not spending another 40% of my time debugging to where this is a net negative. I’m keeping the llm in check on minimizing the slop. - I feel that seasoned engineers, like myself, have a strong advantage over junior developers using Gen ai as a code accelerator because we know generally what the llm outputs and can audit it. Junior engineers that have never spent several hours debugging a seg fault back in the day won’t have the same advantage nor will they understand when the llm is outputting a bunch of slop that can be dramatically simplified - I have worked a little more on the rules aspect of copilot and now when I ask it to go do stuff, it will generally only use my preferred packages to solve the problems e.g. duckdb and spark. It won’t reach for pandas or some obscure library Needless to say, I don’t see myself ever going back to not using a code accelerator anymore. Sure, it would be fun once a quarter to hand role my own code and only rely on a Google search, but I see those days coming to an end. There has been a good quote floating around lately (not sure who coined it) that says: “AI won’t replace software engineers. But the software engineers that use AI will replace the ones that don’t” This, from my experience, I think rings true #dataengineering

218

73 Comments

OpenVSP 3.46.0 Released A few features and a bunch of bug fixes. Good stuff all around, everyone should update unless you're stuck on 3.44.X for some reason. If you are, you should try this out to see if it addresses any of the problems you've had migrating. 3.44.X is not going to receive further updates. https://lnkd.in/gvFsjr7r

71

4 Comments

I didn't learn one important aspect of the principal engineer role until I left my first job with that title. Three different intermediate devs privately told me something along the lines of "I stayed here because you seemed to have faith in the company." My presence in a principal role isn't just about what I can do for the company or what it can do for me. It is a vote of confidence in both the company and in its leadership. Just by being employed I am implicitly telling other employees, "How they treat you here is fine. This is a good company."

94

9 Comments

At what point during civil certification projects is a flying prototype needed? Are flying aircraft actually "needed" in order to establish requirements for certification? Several firms' press releases claim exactly that as they tout their prototype's being approved to fly (i.e., experimental airworthiness certificate). Some clarity here appears necessary. One of my favorite of the 7 Habits is "begin with the end in mind". So I've prepared this graphic to show -- in reverse order (top-down) -- some of the major steps leading to fielding a new aircraft that's used in airline-style operations. Notice that an experimental CofA (permission to fly prototype) is at the far lower left in this diagram (i.e., very, very early in the process). While it is possible that data from a prototype could be used in establishing means of compliance, that does not come from a few success-oriented demonstrations. Instead, they come from organized test programs (with FAA participation) which comprehensively cover off-nominal concerns in addition to nominal performance. A good example of reports from this type of research can be found here: https://lnkd.in/gXBh5xkk Even if a flying prototype is involved with a research project, the result would need to inform means of compliance issue paper(s). Issue papers can take years to conclude. Adding the flying research time only adds to the elapsed time prior to having an accepted certification plan. Afterward, another couple years may be required to conduct the certification project. Conclusion? Having a flying prototype doesn't facilitate the certain end of a certification project until the applicant and the FAA have agreed on what needs to be accomplished WITH that prototype. For that, an accepted certification plan is necessary. Monitoring news for cert plan acceptance by the FAA is far more relevant to tracking schedule progress than is a prototype's first flight. As a flight tester, this pains me to admit. But it's true.

318

67 Comments

Defense innovation rarely fails in engineering; it often fails in authorization. RMF expectations, impact levels, inheritance, continuous monitoring, and evidence requirements stop more transitions than technical limitations ever do. RAPIDFORT, supported by partners such as Carahsoft and SpaceWERX, is offering complimentary enrollment in “From Prototype to ATO,” a 10-week executive cohort taught by Henry Sienkiewicz 金凱旋 (former DISA CIO/DAA). Learn how to interpret RMF, apply reciprocity, streamline validation, and articulate readiness to government stakeholders. 👉 Register: https://lnkd.in/gVDSNtZT #dod #ATO #RMFCompliance #DefenseStartups

49

2 Comments

“But what about NATS?” - I hear this all the time—especially after my recent posts on MQTT. 𝐅𝐢𝐫𝐬𝐭: No single message broker or protocol is a silver bullet. Neither MQTT nor Kafka nor AMQP nor NATS. Never! 𝐒𝐞𝐜𝐨𝐧𝐝—𝐚𝐧𝐝 𝐭𝐡𝐢𝐬 𝐢𝐬 𝐰𝐡𝐲 𝐈’𝐦 𝐰𝐫𝐢𝐭𝐢𝐧𝐠: Synadia, the original donor of NATS, has told the CNCF it plans to withdraw the project, re-license it under the Business Source License (BUSL), and reclaim community assets such as nats.io and the GitHub organization. Industry experts (e.g., Alex Kritikos) have warned me for years that NATS wasn’t truly open source / had open standards. Now I see their point. 🔗 Full CNCF statement: https://lnkd.in/eAwxEdBe

39

12 Comments

Systel’s Raven-Strike II answers the computing demands of C-UAS. Raven is engineered to withstand the worlds harshest environments, bringing data-center-level compute to the tactical edge.  - Modular Open Systems Approach (MOSA) - Standards-based Open Architectures - COTS technologies

61

Just finished reading The C++ Programmer’s Mindset by Sam Morley (generously provided by Mansi S. and Packt), and it is one of the more thoughtful C++ books to come out recently. Instead of starting from syntax and language features, it starts from how you think about problems and then shows how modern C++ helps you execute on that thinking. The first chapter alone sets the tone: decomposition, abstraction, pattern recognition, and algorithm design are treated as core habits, not academic buzzwords. The examples move from “too mathematical” on purpose into more grounded scenarios, which works well if you live in both theory and production code. What stood out most is the explicit link between computational thinking and idiomatic C++: how your choice of abstractions, patterns, and algorithms is shaped by the language and its standard library. The book keeps reminding you that your job is not just “make it compile,” but to design solutions that fit into larger systems, constraints, and teams. Even in early chapters, there is a consistent emphasis on iteration, experimentation, and good practice: small steps, feedback, and testability are treated as first-class tools for problem-solving in C++. This is especially valuable if you already know C++ syntax but feel your design and modeling skills lag behind your implementation skills. If you are an intermediate or advanced C++ developer who wants to sharpen problem-solving, not just memorize more features, this book is worth your time. It reads like a bridge between computer science thinking and day-to-day C++ engineering, and it succeeds at making that bridge feel practical and approachable.

145

15 Comments

🗣️ Our newest Zarf Tech Talk recap is live! 📹 Tech Talk: Simplifying DevSecOps in Air-Gapped Environments with Zarf, brought together experts from Sonatype, Defense Unicorns, and Boeing to break down one of the biggest challenges in secure software delivery: operating in disconnected or restricted environments. Whether you're working in regulated sectors or simply want a more predictable, secure supply chain, this talk is packed with practical lessons. 📺 Watch the recording: https://lnkd.in/e_fNNeQB 📝 Read the full recap: https://lnkd.in/eTukHgMx 📄 Access the slides: https://lnkd.in/eVXkygZS Thanks to our moderator ⚙️ Eddie Knight and speakers Brandt Keller, Kit Patella, and Daniel Miller! 🙏 #Zarf #DevSecOps #OpenSSF #TechTalk #OSSSecurity

25