Research Areas
Domains of inquiry
Six interconnected research domains, unified by a commitment to evidence, reproducibility, and provenance.
Engineering Intelligence
Engineering Intelligence
Engineering Intelligence studies how the knowledge generated during design and verification — failures, root causes, decisions, and rationale — can be captured, connected, and reused. Most of this knowledge is produced once and lost; we research systems that make it a durable, queryable asset.
Research Themes
- Cross-project engineering memory
- Provenance-anchored knowledge graphs
- Root-cause aggregation and retrieval
- Decision and rationale capture
Current Initiatives
- XRecall: a provenance-anchored cross-project engineering-memory architecture
- Reproducible, anti-circular evaluation of engineering-intelligence systems
Future Directions
- Retrieval-quality benchmarks for engineering knowledge
- Validated, fine-grained failure taxonomies
Semiconductor Intelligence
Semiconductor Intelligence
Semiconductor Intelligence applies rigorous, reproducible methods to the analysis of register-transfer-level design and verification. We work on real open-source hardware and independently-derived ground truth rather than synthetic benchmarks.
Research Themes
- RTL design analysis with real elaboration
- Verification failure analysis
- Independent ground truth from maintainer histories
- Reproducibility on real corpora
Current Initiatives
- Evaluation on picorv32, Ibex, CVA6, and OpenTitan corpora
- Maintainer-fix-commit ground-truth construction
Future Directions
- Clock-domain and FSM analysis validated against reference tools
- Industrial regression-data studies
Enterprise Intelligence
Enterprise Intelligence
Enterprise Intelligence extends the principles of provenance and reproducible reasoning beyond hardware to organizational knowledge — connecting decisions, evidence, and outcomes so that institutional knowledge survives turnover.
Research Themes
- Institutional memory
- Evidence-linked decision support
- Provenance across knowledge work
Current Initiatives
- Foundational architecture for provenance-anchored organizational memory
Future Directions
- Cross-domain knowledge graphs
- Auditable decision trails
Adaptive Computing
Adaptive Computing
Adaptive Computing studies optimization and computation methods that adapt to problem structure, evaluated honestly against established solvers and baselines, including the regimes where they do and do not help.
Research Themes
- QUBO and Ising problem formulations
- Classical and metaheuristic solvers
- Honest baseline comparison
Current Initiatives
- Comparative studies against established constraint solvers
- Documented extension points for future quantum backends
Future Directions
- Regime-aware solver selection
- Reproducible optimization benchmarks
Memory-Centric Computing
Memory-Centric Computing
Memory-Centric Computing investigates architectures in which memory, lineage, and provenance are first-class. Our kernels reduce computation to canonical, signed inputs so that results are deterministic and byte-reproducible.
Research Themes
- Signature-based determinism
- Lineage and propagation primitives
- Idempotent ingestion
Current Initiatives
- A reusable intelligence kernel shared across applications
- Byte-reproducibility test discipline
Future Directions
- Formalized provenance semantics
- Reuse measurement methodologies
Authenticity Intelligence
Authenticity Intelligence
Authenticity Intelligence researches methods to establish the origin, integrity, and authenticity of artifacts — from engineering results to digital records — using cryptographic signatures and traceable provenance chains.
Research Themes
- Cryptographic provenance
- Artifact integrity and traceability
- Verifiable reproducibility
Current Initiatives
- SHA-256 signature chains for reproducible results
Future Directions
- End-to-end authenticity for distributed engineering workflows