AI Accelerator Silicon — Microsoft
Verify floating-point pipelines and coverage closure for next-gen AI chips.
Built reusable UVM environments and automation adopted across the org.
Silicon Engineer · Founder
Ernesto Sola-Thomas
Building Huitzo — an AI intelligence platform for enterprise. Microsoft silicon engineer with the technical depth to make it real.
- Founder of Huitzo: intelligence-as-a-service that gives organizations real AI depth, not demos.
- Silicon verification engineer at Microsoft — AI chip infrastructure, formal verification, and systems depth.
Building Now
Huitzo
An intelligence platform turning advanced AI into enterprise outcomes. Early-stage, serious architecture.
Explore Profile
Track Record
Selected Proof
High-signal work across silicon, AI, research, and venture building.
Huitzo — Intelligence as a Service
Make advanced AI capabilities accessible to enterprises and startups.
Designing scalable AI infrastructure with modular data pipelines.
Learn morePost-Quantum Secure File Transfer
Protect file transfers against quantum computing threats.
Published a PQ-secured protocol with blockchain auditability (arXiv).
Learn moreAutonomous Wheelchair Platform
Enable hands-free indoor navigation for mobility-impaired users.
Published at IEEE — low-cost, open-source design with dual control modes.
Learn moreCloud-Native Platform Ops
Ship production-grade billing and auth for multi-tenant SaaS.
Automated CI/CD pipelines enabling weekly iteration with confidence.
Highlights
Achievements & Focus Areas
A snapshot of how engineering, research, and entrepreneurship intersect across my work.
Silicon verification
Building reusable UVM test environments for RTL at Microsoft, covering floating-point pipelines, constrained-random testing, and coverage automation.
AI & hardware systems
Verification of AI accelerator silicon and chip-level architectures spanning formal methods, functional coverage, and synthesis-ready design.
Cloud infrastructure
Deploying production-grade platforms with Kubernetes, Docker, Terraform, and CI/CD pipelines for personal ventures and client engagements.
Applied research
Post-quantum cryptography, assistive robotics, and neurotech — translating academic research into working prototypes.
Journey
Professional Experience
From AI silicon verification to research and consulting, I bridge deep engineering with purposeful outcomes.
- Founder & Principal Engineer 2025 — Present
Sola-Thomas Solutions LLC
Remote
Consulting studio helping founders and small teams de-risk architecture, ship prototypes, and build production-grade infrastructure.
- Building Huitzo, an intelligence-as-a-service platform for enterprise and startup clients.
- Delivered 15+ client engagements across fintech, logistics, healthcare, and consumer tech.
- Shipped infrastructure patterns adopted by venture-backed startups for scale and compliance.
- Available for consulting engagements online and in the Raleigh-Durham area.
Cloud architectureHardware-software prototypingAnalytics platformsCI/CD automation - Silicon Engineer 2022 — Present
Microsoft
Raleigh, NC
Verifying AI acceleration silicon through UVM environments, automation, and cross-team collaboration.
- Led floating-point pipeline verification for next-gen AI accelerator, ensuring RTL correctness across corner cases.
- Built automation harnesses that accelerated regression cycles and reduced manual debug hours across verification teams.
- Partnered with design and architecture to close coverage gaps using constrained-random scenarios and data-driven insights.
- Owned implementation and maintenance of formal and functional testbenches.
- Built scalable, reusable tools that were adopted across the organization.
SystemVerilogPythonConstrained random verificationCI orchestration - Undergraduate & Graduate Researcher, Teaching Assistant 2020 — 2025
Clarkson University
Potsdam, NY
Investigated resilient distributed systems blending embedded hardware, software, blockchain, and post-quantum cryptography while collaborating across disciplines.
- Designed a post-quantum secure file transfer protocol with blockchain auditability for my master's thesis (see arXiv:2504.07938).
- Built embedded systems projects using FPGAs, microcontrollers, custom PCBs, and 3D-printed enclosures in collaboration with students and professors.
- Collaborated with cross-disciplinary teams exploring AI, quantum fundamentals, and trusted computing deployments.
PythonSystemVerilogC/C++FPGALinux
Engagement
Ways I Can Help
Outcome-driven engagements for founders, teams, and organizations building at the edge of what's possible.
Technical due diligence
Evaluate architecture, code quality, and scalability risks for investment decisions or acquisition targets.
Prototype & system architecture
Design and build proof-of-concept systems spanning hardware, firmware, cloud, and AI pipelines.
AI + hardware systems strategy
Advise on AI accelerator integration, silicon-software co-design, and intelligent infrastructure planning.
Product & engineering advisory
Help early-stage teams make high-leverage technical decisions — from stack selection to team structure.
Early-stage technical direction
Partner with founders to translate vision into architecture, roadmap, and first working system.
Start a conversation
Select engagements only. All advisory work is independent of my role at Microsoft.
Projects
Making ideas tangible
A list of past projects ranging from robotics, secure compute, simulation tooling, and cloud platforms.
Software & Platforms
2 initiativesMonte Carlo GPU Finance Simulator
2025
Built GPU-accelerated Monte Carlo simulations in Python to model portfolio risk and derivative pricing scenarios.
- Delivered 40x speed-ups vs. CPU baselines for client prototypes.
- Equipped stakeholders with faster decision support dashboards.
PythonCUDANumPy
The work combined quantitative methods, GPU programming, and data storytelling to turn raw simulation output into actionable strategy.
Cloud-Native Platform Ops
2024
Orchestrated Django, Supabase, and Stripe ecosystems with Dockerized deployment across Azure Kubernetes Service and Cloudflare.
- Delivered reliable billing and auth flows for multi-tenant applications.
- Automated build and release pipelines, enabling weekly iteration with confidence.
DjangoSupabaseStripeAzure Kubernetes ServiceDocker
This stack demonstrates how I build for resilience—pairing infrastructure automation with thoughtful developer experience.
Research & Exploration
3 initiativesAI Chip Verification Suite
2025
Building reusable verification environments and analytics to accelerate AI accelerator bring-up and coverage closure.
- Reduced test debug time through targeted stimulus and health dashboards.
- Improved confidence in floating point paths for production delivery.
SystemVerilogPython automationMetrics pipelines
A focused initiative extending core verification work into reusable frameworks and knowledge transfer artifacts for future silicon programs.
ECG-Based Emotion Recognition
2024
Enhanced emotion recognition accuracy using wearable electrocardiogram (ECG) data and machine learning models to classify emotions across valence, arousal, and dominance dimensions.
- Demonstrated superior performance of random forests with carefully selected ECG features for emotion classification.
- Advanced feature engineering techniques for physiological emotion recognition systems.
- Contributed to personalized mental health monitoring and human-computer interaction applications.
- Established framework for real-time emotion detection using wearable ECG devices.
PythonMachine learningECG signal processingRandom forestsSupport vector machines (SVM)Logistic regressionDREAMER datasetFeature engineering
FPGA Vision Processing for Prosthetic Hands
2023
Implemented vision-based automated control for prosthetic hands using FPGA processors, featuring an RGB camera interfaced to Zynq UltraScale+ MPSoC with object detection deep learning models.
- Addressed high-power consumption limitations in sophisticated robotics through power-efficient FPGA implementation.
- Developed advanced image processing pipeline from scratch for prosthetic hand applications.
- Created publicly available repository with thorough implementation documentation.
- Enabled vision sensors to artificially sense objects from wrist position for sophisticated grabbing control.
Zynq UltraScale+ MPSoCFPGA designComputer visionDeep learningImage processing pipelineRGB camera interfaceDigital circuit design
Video Archive
Projects in Motion
Recorded demos and documentation of past hardware and research projects — a window into earlier engineering work.
Autonomous Wheelchair Platform
2021
A low-cost, open-source autonomous wheelchair built on a standard manual chair. Demonstrated hands-free indoor navigation using stereo depth camera, LiDAR, and Nvidia Jetson. Published at IEEE.
Watch on YouTube
NBI EMG Prosthetic Hand
2021
A prosthetic hand driven by EMG muscle signals. The system translates electrical muscle activity into multi-grip actuation, restoring confidence and capability through biofeedback.
Watch on YouTube
Schumilop Water Quality Monitor
2019
A science fair project born from discovering algae in the school water supply. Combined electrolysis and sensor electronics to build a working water quality detection instrument.
Watch on YouTube Collaborate
Let's build what's next
Whether you're shipping silicon, scaling a venture, or exploring secure systems, I'd love to hear about the challenge.
Primary contact
[email protected]What would you like to discuss?
Location
Raleigh, NC
Current availability
Open for select advisory, prototyping engagements, and speaking with founders shaping technology.