Currently: Product engineering, backend systems, and cloud infrastructure
Focus: Memory management, distributed systems, performance engineering
Building toward: Deep systems work and research roles

Allocators using coarse-grained reclamation (slabs, arenas, epochs) can exhibit unbounded retention even when all objects are freed. A single long-lived allocation pins an entire granule. This work formalizes when bounded retention is possible and provides measurement tools validated on Redis, where deleting 50% of keys freed 195K objects but reclaimed 0 slabs

→ Explore the project

Components: Paper • Theory • Measurement • Reference Implementation • Validation

Event-driven systems can converge without coordination even when operations violate invariants and don't commute—if compensation is well-founded and commutative. This work identifies a third convergence regime alongside CRDTs and invariant confluence, where compensated results commute even though operations don't.

→ Explore the project

• WFC + CC conditions guarantee convergence via Newman's Lemma
• Verification calculus reduces global CC obligation to local per-event-pair checks
• Build-time verification tools with exhaustive state-space enumeration

Components: Paper • Landing Page • gsm Library • nccheck Verifier

Extends normalization confluence to federated environments where multiple registries are connected by morphisms encoding cross-organizational constraints. For tree-shaped networks, proves federated convergence requires only morphism validity preservation—all other conditions derive from network acyclicity via an authority argument

• Authority argument: Source normal forms deterministically control target shared components
• Federated construction: Reduces multi-registry problem to single-registry case
• Necessity proofs: Acyclicity required (cyclic morphisms cause compensation cycles), validity preservation required (morphism repair and local compensation oscillate)

Components: Core Paper • Federated Paper • gsm Library • nccheck Verifier

Languages:

Platforms & Shells:

Developer Tooling:

Artificial Intelligence:

• Product Engineer
• Cloud/Platform & backend engineer: APIs, Infrastructure as Code, Automation
• Compliance Automation for FDA-regulated systems
• Exploring blockchain security and on-chain systems
• Building developer tooling: secrets orchestration, reproducible environments, and CLI workflows