Saphira AI (YC S24)

𝗘𝘃𝗲𝗿𝘆 𝗻𝗲𝘄 𝘃𝘂𝗹𝗻𝗲𝗿𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝘀𝗵𝗼𝘂𝗹𝗱 𝘁𝗿𝗶𝗴𝗴𝗲𝗿 𝗮 𝗻𝗲𝘄 𝗧𝗔𝗥𝗔 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗰𝗮𝗹𝗹𝘆. That is the shift VicOne and Saphira AI (YC S24) are operationalizing. Turning vulnerability signals into 𝗟𝗶𝘃𝗶𝗻𝗴 𝗧𝗔𝗥𝗔. A continuously evolving risk model that stays aligned with real-world threats. When a new vulnerability is identified, TARA is no longer static documentation. 𝗜𝘁 𝗯𝗲𝗰𝗼𝗺𝗲𝘀 𝗮 𝗹𝗶𝘃𝗶𝗻𝗴 𝘀𝘆𝘀𝘁𝗲𝗺. Updating impacted assets, reassessing attack paths, and generating mitigation actions with real engineering context. This closes a persistent gap. Where vulnerability management operates continuously, but cybersecurity documentation lags behind actual risk. The result is not just automation. It is a continuous cybersecurity loop where risk assessment, response, and compliance evidence evolve together. Supporting audit readiness while improving prioritization. What’s your biggest challenge today: detecting vulnerabilities, or keeping your risk assessments continuously aligned with them? Learn more about the partnership: https://lnkd.in/gya_Dprx #AutomotiveCybersecurity #CyberRisk #SDV #ThreatIntelligence #ISO21434 #ConnectedVehicles #PhysicalAI

𝗘𝘃𝗲𝗿𝘆 𝗻𝗲𝘄 𝘃𝘂𝗹𝗻𝗲𝗿𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝘀𝗵𝗼𝘂𝗹𝗱 𝘁𝗿𝗶𝗴𝗴𝗲𝗿 𝗮 𝗻𝗲𝘄 𝗧𝗔𝗥𝗔 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗰𝗮𝗹𝗹𝘆. That is the shift VicOne and Saphira AI (YC S24) are operationalizing. Turning vulnerability signals into 𝗟𝗶𝘃𝗶𝗻𝗴 𝗧𝗔𝗥𝗔. A continuously evolving risk model that stays aligned with real-world threats. When a new vulnerability is identified, TARA is no longer static documentation. 𝗜𝘁 𝗯𝗲𝗰𝗼𝗺𝗲𝘀 𝗮 𝗹𝗶𝘃𝗶𝗻𝗴 𝘀𝘆𝘀𝘁𝗲𝗺. Updating impacted assets, reassessing attack paths, and generating mitigation actions with real engineering context. This closes a persistent gap. Where vulnerability management operates continuously, but cybersecurity documentation lags behind actual risk. The result is not just automation. It is a continuous cybersecurity loop where risk assessment, response, and compliance evidence evolve together. Supporting audit readiness while improving prioritization. What’s your biggest challenge today: detecting vulnerabilities, or keeping your risk assessments continuously aligned with them? Learn more about the partnership: https://lnkd.in/gya_Dprx #AutomotiveCybersecurity #CyberRisk #SDV #ThreatIntelligence #ISO21434 #ConnectedVehicles #PhysicalAI

Join the #IEEE #VTS & #PSES in‑person technical forum: “The Intelligence of the Machine. The Rigor of the Road.” May 12, 2026 | 5:00 – 7:30 PM PDT SEMI, 673 S Milpitas Blvd, Milpitas, CA Leading experts will cover automotive systems, AI, firmware, and safety. Speakers: François Piednoel de Normandie (Athos Silicon ) – Chiplet architecture & fault-tolerant systems Sachin Athanikar (Mercedes-Benz Research & Development North America, Inc.) – Automotive software, UCIe & chiplets, ADAS Dong Wei, MBA (Arm) – UEFI & firmware standards, UCIe / CXL Prof. Ding Zhao (Carnegie Mellon University) – Safe AI, rare‑event safety, reinforcement & imitation learning Akshay Chalana ( Saphira AI (YC S24) – Continuous risk & compliance for SDVs Why attend: Learn how industry leaders standardize automotive firmware for safety & security Gain practical insights into AI-driven vehicle systems Network with top experts and peers 🎯 Open to industry professionals & students — seats are limited! 👉 Register now: https://lnkd.in/gVRrzbdE #IEEE #VTS #ADAS #AI #SDV #Networking #Innovation Co sponser -  IEEE Student Branch SFBU & IEEE SCV Affinity Group T.Kim Parnell Lingqiao Qin Dr. Anantharaj Thalaimalai Vanaraj Siva Uppalapati Dhaval Brahmbhatt Aishwarya Ashok Patil Aditi Ramadwar Abhishek Vyas Zac Lochner, PhD, CFEI Amogh Jayavant Silky Sindhani IEEE Santa Clara Valley Section

Thrilled to join the NVIDIA Halos AI Systems Inspection Lab! Excited to collaborate on advancing trustworthy AI for autonomous systems with NVIDIA’s ecosystem leadership. Read the full announcement here: https://lnkd.in/gcmfYf46 #GTC2026 #AISafety #AutonomousVehicles #ISO26262

I’m very excited to share that the pre-print of our new paper is now available: “ML FMEA in Action: Lessons from Applications of Machine Learning Safety” You can read it here: https://lnkd.in/erGCBSgT This one is especially meaningful to me because it is one of my first publications, and I’m grateful that it’s alongside such an outstanding group of co-authors. As machine learning is increasingly deployed in safety-critical systems, we need methods that account for the realities of learning-based development. Traditional safety approaches were built around deterministic software, but ML systems depend on data, statistical behavior, and evolving operating conditions, which makes them harder to predict, verify, and assure. In this paper, we share lessons from applying the Machine Learning Failure Mode and Effects Analysis (ML FMEA) method in real engineering programs, including examples from autonomous vehicles and humanoid robots. A central goal of the work is to connect ML development decisions more directly to system-level safety outcomes. What makes this paper particularly exciting is that it represents a truly collaborative effort: 13 authors across 8 organizations and multiple industries, all working toward safer deployment of machine learning in critical applications. I’m especially thankful to my esteemed co-authors, whom I’m proud to be associated with on this work: Paul Schmitt, Neil Wadhvana, Bodo Seifert, David Ward, Majed Mohammed, Michael Wagner, Simon Diemert, Fahim Mannan, Jerry Lopez, Krzysztof Pennar, Chaitanya Shinde, and Justin Poh. I’m also excited that the work is open source, which I hope helps make these ideas more useful and accessible to the broader community. We’ll be presenting this at SAE World Congress (WCX) 2026 in Detroit on April 15, and I’d be thrilled to connect with others interested in ML safety there. Many thanks as well to Saphira AI (YC S24), Reynolds & Moore, TÜV Rheinland Group, Aptiv, Edge Case, HORIBA MIRA, Critical Systems Labs, Torc Robotics, and the reviewers at SAE International World Congress.

🚨 We’re heading to GTC — and hosting a Safety Dinner 🚨 AI and robotics capabilities are accelerating fast. The question we hear everywhere is the same: How do we deploy these systems safely in the real world? During NVIDIA GTC, Saphira AI (YC S24) is hosting an evening dinner for engineers, founders, and safety leaders working on: * robotics & autonomy * AI infrastructure * EHS risk assessment * electronics product compliance * functional safety & certification We’ll be discussing the practical side of deploying safe, certifiable AI systems, including topics like: • robotics safety standards (ISO 10218, ISO 13849) • automotive safety (ISO 26262) • environmental health & safety risk workflows • product certification for electronics & robotics • scaling compliance across rapidly evolving systems At Saphira we’re building AI tools that automate compliance workflows — helping teams turn complex standards into structured, traceable requirements and certification artifacts. If you're attending GTC, come join us for dinner and conversation with others working on the future of safe AI and robotics. 📍 Location shared with confirmed guests 🍽 Dinner & drinks 🤝 Great safety conversations Register here: https://luma.com/ywcr654i

Most safety audits don’t fail because engineers didn’t do the work. They fail because the evidence isn’t consistent or complete. AI is starting to change that — but only when used correctly. In 3 days, our CEO, Akshay Chalana, will be speaking with TÜV Rheinland North America’s Bodo Seifert about how AI fits within established functional safety standards — including IEC 61508, ISO 26262, and ISO/IEC TR 5469. We’ll cover: ✔ How tool qualification applies to AI ✔ Where AI tools are high-risk (and where they aren’t) ✔ What “TCL1” really means in practice ✔ How AI can support — not replace — human judgment If you’re working in robotics, automotive, or safety-critical electronics, this will be a practical session. 📅 Feb. 18, 2026 ⏰ 10 AM CST | 8 AM PST | 5 PM CET Register ➡ https://gag.gl/Sjmptm #FunctionalSafety #AI #SafetyCritical #EngineeringLeadership #ISO26262

🚀 Yes, you’re seeing this right - we’re on the Automate Show 2026 together! Two teams. Two continents. One shared vision. Saphira AI (YC S24) and Müller & Partner Sachverständige • The Safety Manufacture are coming together to connect 🤖 Safety AI from San Francisco with 🦺 real, hands-on physical safety expertise from Southern Germany. This isn’t about theory. It’s not about slides or buzzwords. It’s about rethinking how safety in automation should feel, work, and scale. As systems grow more complex, one thing becomes clear: Safety solutions must become simpler. 📍 Automate Show 2026 📍 Booth #4075 👉 Stop by and experience it live. See for yourself what becomes possible when Safety AI meets real-world safety engineering - across continents, cultures, and disciplines. Let’s talk. Let’s challenge the status quo. Let’s shape the future of safety together. Jens Müller, CMSE®

It was a pleasure to present at Volvo Group’s CampX GoWest Side Event in Gothenburg yesterday!

Check out what we’re doing at STS with the help of our partner Saphira AI (YC S24)