I help engineering teams turn complex C/C++, embedded, AI, and research challenges into verified, auditable deliverables. Services include C/C++ debugging and modernization, MISRA and safety compliance, SBOM and supply-chain security, AI-tool assessment, Wolfram and Lean formal verification, and embedded Linux or accelerator prototyping.
Portfolio: https://zuwasi.github.io/Public-html-pages/
Each service below links to a public presentation, demonstration, analysis, or implementation that shows how I approach the work. These examples are a fraction of a larger portfolio, much of which consists of confidential, bespoke customer projects.
Reproduce difficult failures, separate symptoms from causes, collect GDB and sanitizer evidence, correct ownership and lifetime defects, and preserve the result with regression tests.
View the C++ debugging evidence lab →Analyze safety-critical code, triage violations, implement controlled fixes, document formal deviations, and verify changes through unit tests and compliance reports.
View the MISRA C:2023 remediation case →Design guided BSP workflows covering component selection, isolated builds, simulation, testing, CVE triage, SBOM, VEX, and evidence-based release gates.
View the Yocto BSP Studio concept →Generate and assess SBOMs, investigate dependency exposure, triage CVEs and malware risk, map licenses, and produce audit-ready evidence for connected or air-gapped environments.
View real-time supply-chain detection →Evaluate AI development tools for enterprise adoption, data handling, access boundaries, supply-chain exposure, governance, and regulated-environment readiness.
View the CISO assessment →Translate research claims into executable models, cross-check numerical behavior, derive invariants, and create kernel-checked proofs for critical properties.
View certified numerical verification →Move from papers and mathematical models to reproducible computation, formal properties, HDL simulation, FPGA-oriented prototypes, and an explicit physical-validation plan.
View the FHE accelerator workflow →Assess CUDA inference pipelines, export compatible models, map preprocessing and operators, build target configurations, and validate performance on edge-AI platforms.
View the CUDA-to-Axelera guide →
MantiQ Factor is my product vision and engineering workflow for converting research and technical documents into traceable, formally verified executable systems. It combines structured document conversion, AI orchestration, Wolfram computation, Lean 4 proof, domain skills, curated knowledge, and code or hardware synthesis.
The goal is not automatic code generation without controls. Every stage is designed to remain inspectable, reproducible, and connected to its source evidence.
ESL AI LogicLab develops local and air-gap-capable applications for environments where code privacy, correctness, traceability, and auditability matter.
Multi-architecture static analysis for assembly language, with rule-based inspection, reviewed AI fix suggestions, and regulatory reporting.
Explore Assemblomator →An isolated local-LLM workflow for precise, reviewable fixes to C, C++, and C# static-analysis findings without sending source code to the cloud.
Explore FixCreator →A coordinated Python analysis workflow combining seven engines with deduplication, unified severity, source context, GUI review, and repeatable CLI execution.
Explore PythonStator →Local React analysis for component health, dependency risk, dead code, modernization planning, security review, and QA quality gates.
Explore ReactStator →Air-gap-ready SBOM, vulnerability, malware, license, AI/ML inventory, and dataset-provenance analysis for regulated software supply chains.
Explore SBOMator →TypeScript-native, air-gapped SAST that correlates multiple scanners, assesses reachability locally, and keeps every proposed fix under human review.
Explore TSscrutinizer →My work spans approximately 250 public and private repositories. More than 120 are publicly visible; most private repositories contain bespoke customer work that cannot be published because it was created for specific organizations and environments.
The public portfolio connects presentations, demonstrations, technical analyses, reports, prototypes, and reproducible engineering workflows across embedded development, AI, security, formal methods, and scientific computing.
I publish reusable demos, educational labs, technical guides, security research, verification workflows, and working source code as a practical donation of engineering time and knowledge to the open-source community.
These public artifacts are intended to help other engineers learn, reproduce results, challenge assumptions, and build safer systems. They are also transparent evidence of the services I can deliver.
Bring the defect, research paper, compliance target, legacy codebase, supply-chain question, or prototype objective. We will define the evidence required and build the smallest credible path to a verified deliverable.