"What I cannot create, I do not understand." — Richard P. Feynman
"The first principle is that you must not fool yourself — and you are the easiest person to fool." — Richard P. Feynman
"The introduction of the cipher 0 or the group concept was general nonsense too, and mathematics was more or less stagnating for thousands of years because nobody was around to take such childish steps." — Alexander Grothendieck
"We can only see a short distance ahead, but we can see plenty there that needs to be done." — Alan Turing
I build high-performance distributed systems at the intersection of AI inference, cryptography, and fault-tolerant computing. I previously co-founded Atoma, building infrastructure for secure GenAI workloads with confidential computing — fast serverless LLM inference with verifiable privacy guarantees.
PhD in algebraic geometry and algebraic topology, working under Bertrand Toen on derived arithmetic geometry. Research focused on derived non-archimedean analytic geometry, including derived formal schemes, Raynaud localization, and non-archimedean Hodge filtrations. Selected publications on arXiv.
- Large-scale LLM inference in Rust — High-throughput GPU serving systems, kernel-level optimization, distributed inference orchestration
- Consensus protocols — Implementing Minimmit and exploring next-gen BFT designs for the Hellas network
- High-frequency low-latency trading systems — Performance-critical infrastructure, lock-free data structures, kernel-bypass networking
- Large-scale LLM inference — High-throughput serving systems, GPU kernel optimization, distributed inference orchestration
- Distributed systems — Consensus protocols, fault tolerance, state machine replication
- Cryptography — Zero-knowledge proof systems (PLONK, Plonky2/3, STARKs), applied cryptographic protocols, verifiable computation
- Confidential computing — TEE-based inference, hardware-rooted trust, secure multi-party AI workloads
- High-frequency low-latency trading systems — Performance-critical infrastructure, lock-free data structures, kernel-bypass networking
🚀 AI Inference
- atoma-infer — Fast serverless LLM inference engine, written in Rust
- atoma-node — Core infrastructure for confidential computing AI inference
- atoma-proxy — Atoma's proxy service for routing and load balancing inference workloads
- nvtrust — Open source tooling to support confidential computing on NVIDIA GPUs
🌐 Distributed Systems
- hellas consensus — Core consensus protocol for the Hellas blockchain, including Minimmit implementation
- ganishka — Distributed systems challenge solutions (Maelstrom) in Rust
- cdks — Context-driven knowledge service ingestion in Rust
🔐 Cryptography & Zero Knowledge
- zktree — A Rust implementation of zero-knowledge Merkle trees
- ProvableMerkleTrees — Recursive Plonky2 circuits for provable Merkle tree operations
- chaum-pedersen-auth — Chaum-Pedersen zero-knowledge authentication protocol in Rust
📐 Research
- Derived_rig_cohomology — Derived rigid cohomology for singular varieties over perfect fields of positive characteristic
- Derived O_k-adic geometry and derived Raynaud localization theorem
- Spreading out the Hodge filtration in non-archimedean geometry
- Derived non-archimedean analytic Hilbert space (with Mauro Porta)
Active contributor to projects across the distributed systems, cryptography, and AI inference ecosystems — including NVIDIA Dynamo, Tari protocol, Plonky3, vLLM, SGLang, raft-rs, reth, Aeron, and various ZK proof system libraries.
| [email protected] | |
| Email (deprecated) | [email protected] |
| Jorge António | |
| X | @Twen2Ack |





