I like machines close to the metal and results you can measure. By day I build in-memory data systems at AWS ElastiCache; the rest of the time I train and compress small fast models, write CUDA, and maintain open-source tools people actually run.

Research notes and working papers live at avifenesh.ai. Issues, questions, and counterexamples are always welcome.

• memra — from-scratch Rust + CUDA inference engine for one Blackwell GPU (formerly bw24). Bit-exact by construction; runs four model families above llama.cpp on the same rig, plain and speculative decode both. v0.69 shipped to crates.io with prebuilt binaries and FP8 checkpoint serving at 2.6–2.8× plain decode.
• hqmtp — MTP draft-head lab, concluded. Function cuts (pruning, low-rank, distillation) pay a 10–19-point off-distribution tax that fidelity cuts don't; the zero-training trimmed-vocabulary recipe won at 1.8–2.7× end to end. The negative results stay in the ledger.
• recipe-lab — layer-loop weight sharing + ε=λ/(N√L) residual scaling, combined for the first time and tested from zero in 11 pre-registered rounds. In the data-constrained regime the looped model beat FLOPs-matched vanilla in all three mixer families — attention, pure SSM, and hybrid; seven paired runs, zero sign flips — with 26–34% fewer parameters. Rule isolated: loop the state-mixer, never the retriever.
• mem-retrofit — grafted a product-key memory layer onto a stock dense 4B and ran it against LoRA over sequential updates. The retrofit is free at lr/10; the published forgetting advantage failed 12/12 confidence intervals.
• More studies with receipts: fixed-compute-frontier (a preregistered kill-gate ledger, ~84 theory lanes) · gemma-expert-atlas (26B MoE expert surgery, 3,840 experts traced) · block-routed-swiglu (near-free kernel, refuted capability) · moe-lab · assumption-excavator.
• Working papers: small-vocabulary MTP heads · prune, heal, quantize — methods, failed arms, and evidence in the open.
• In review upstream: NVFP4 quantization at llama.cpp, hybrid-KV loads at vLLM and LMCache.

• Valkey GLIDE — the official multi-language Valkey/Redis client. Rust core, Java/JNI, Node/N-API; 158 PRs authored, 480 reviewed, plus releases and the support that starts after the code works.
• valkey-skills — official AI skills for the Valkey ecosystem. Started it, maintain it.
• agent-sh — my org: an ecosystem of tools for agent-assisted development, working across Claude Code, Codex, OpenCode, Cursor, and Kiro.
• glide-mq — Node.js queue on Valkey Streams with a Rust N-API core, plus adapters for Hono, Fastify, Hapi, NestJS and a dashboard.
• Also around: RustOwl (runtime, memory, and CI work) and ChatGPT Desktop for Linux (Wayland input, accessibility, packaging).

• Valkey — core contributor; sync-from-replica replication in review. On the side: CRIU copy-on-write live-migration research — under 50 ms of freeze while migrating a 200 GB loaded instance.
• FlowFabric — durable-execution engine in Rust for Valkey, Postgres, and SQLite: lease-safe workers, waitpoints, budgets.
• layout-audit — DWARF memory-layout analysis: padding, layout diffs, size budgets for C/C++/Rust/Go.
• scrump — format-aware secret scrubber for binary capture artifacts: perf.data, core dumps, nsys traces, JFR.
• ocaml-valkey — OCaml 5 + Eio Valkey client, published on opam.

• agnix — linter and language server for AI agent configs: 444 rules with autofixes, a GitHub Action, an MCP server, and editor plugins.
• computer-use-linux / agent-workspace-linux — Linux desktop control over MCP, and isolated agent-owned desktops so an agent never has to touch your real machine.
• eigen — local agent daemon in Go: resumable sessions, multi-provider routing, memory, observability.

• Talks: Inside Valkey GLIDE on the AWS Developers Podcast · Glide into resiliency on Let's Talk About Data
• Writing: avifenesh.ai/writing · answering on Stack Overflow
• 📫 aviarchi1994@gmail.com · LinkedIn · X

If something here saved you time, sponsoring helps me keep doing it.