David Fiala - Chief Scientist, Architect of Deterministic AI
Logital’s Chief Scientist & Founder, David Fiala, brings a lifetime of forging trust in the world’s most complex and critical computational systems. His career represents an exceptionally rare convergence of deep supercomputing resilience, elite planetary-scale security architecture, and a foundational commitment to verifiable outcomes-the precise combination required to make AI truly deterministic and reliable.
This isn’t just experience; it’s a history of applying a fanatical level of precision to systems where a single overlooked detail could have catastrophic consequences. Logital AI’s deterministic inference is a direct result of this singular expertise, embedding uncompromising rigor into every layer of execution.
The Foundation for AI Reliability
Section titled “The Foundation for AI Reliability”David Fiala’s journey has been defined by an uncompromising mindset toward computational integrity and reliability:
-
Supercomputing & Foundational Resilience: His PhD at North Carolina State University wasn’t just academic; it involved developing foundational algorithms and compiler-level innovations (e.g., CUDA ILP backends) that protect the world’s largest supercomputers in collaboration with Oak Ridge and Sandia National Laboratories. This work tackled silent data corruption- where a single flipped bit can invalidate nation-critical research, instilling a cellular level understanding of computational integrity.
-
Planetary-Scale Security at Google:
- In Google’s Central Privacy & Security, he co-founded the internal “root” production security team, redesigning privileged access for 85,000 engineers and uncovered sophisticated insider attacks that had operated undetected for years.
- As the founding head of security for Google’s AI healthcare initiative, he architected the frameworks for compliant, secure storage and processing of millions of Americans’ medical records, navigating extreme regulatory scrutiny.
-
Patented Innovation in System Verifiability: His patents in areas like pre-identifying malicious rootkit behavior and real-time behavioral analysis underscore a proactive, forensic approach to system integrity.
Core Credentials
Section titled “Core Credentials”- PhD, Computer Science, North Carolina State University
- Dissertation: “Transparent Resilience Across the Entire Software Stack for HPC Applications”
- M.S. & B.S., Computer Science, Middle Tennessee State University
- Key Previous Roles:
- Head of Security, Google Health AI
- Google Central Privacy & Security (co-founder, “root” production security team; Infrastructure Exploit Research)
- High-Performance Computing Research Leader (NCSU, with Oak Ridge & Sandia National Labs)
- Select Publications:
- Detection and Correction of Silent Data Corruption for Large-Scale High-Performance Computing: International Conference for High Performance Computing, Networking, Storage and Analysis
- Mini-Ckpts: Surviving OS Failures in Persistent Memory: 30th ACM International Conference on Supercomputing
- FlipSphere: A Software-based DRAM Error Detection and Correction Library for HPC: 20th IEEE/ACM International Symposium on Distributed Simulation and Real Time Applications
- Select Patents:
- US 9,323,929 B2: Pre-identifying probable malicious rootkit behavior using behavioral contracts
- US 9,158,604 B1: Lightweight data-flow tracker for real-time behavioral analysis using control flow
- Y Combinator Alumnus (W23) with Teclada Inc. (Logital AI)
Build on Certainty
Section titled “Build on Certainty”Partner with Logital AI. Leverage deterministic AI architected by a leader who has mastered reliability where the stakes are highest.