HorneroConfig
Comprehensive dotfiles framework for any WM/OS — reproducible, layered, and documented like a nest built stick by stick.
Explore dotfilesDigital Nest · Warm, precise, reproducible
Build your nest from scratch
I'm Ulises Jeremias — Solutions Architect at NaNLABS, Core Team Member at V, and AUR Maintainer. I build developer tooling, AI-powered workflows, CLI apps, and infrastructure automation — mostly in TypeScript, Go, Shell, and Python.
I help teams ship reliable tooling: from workstation foundations (HorneroConfig) and AI workspaces to scientific libraries for the V language and community-driven CLIs.
Solutions Architect, V Core Team, AUR Maintainer
Warm, precise, and reproducible — I help teams ship tooling that survives reinstalls and handoffs.
Ulises JeremiasHe/Him
Solutions Architect @ NaNLABS
La Plata, Buenos Aires, Argentina
Solutions Architect at NaNLABS, Core Team Member at V, and AUR Maintainer. I build developer tooling, AI-powered workflows, CLI apps, and infrastructure automation — mostly in TypeScript, Go, Shell, and Python.
- Solutions Architect · NaNLABS
- Core Team Member · V Language
- AUR Maintainer · Arch User Repository
- Open Source Enthusiast
I help teams ship reliable tooling: from workstation foundations (HorneroConfig) and AI workspaces to scientific libraries for the V language and community-driven CLIs.
What I’m building now
Active, not aspirational — each tile is a repo you can read, run, and critique today.
Agentic Workstation
Personal AI developer workstation — agents, skills & CLI helpers that make any assistant useful on day one.
Explore workstationAgentic Harness
Portable workspace orchestrator — persistent memory, personas, skills for Claude Code, Cursor, opencode, and Gemini CLI.
Explore harnessV Scientific Library (VSL)
Scientific computing for V — numerical methods, linear algebra, and statistics with precise, tested implementations.
Explore VSLV Tensor Library (VTL)
Tensor and autograd for V — building ML primitives with clarity over magic.
Explore VTLA universe of worlds, woven together
Each world is a documentation-ready section — dotfiles, workstation, toolkit, V, templates, and community — connected, not scattered.
dotfiles · workstation
HorneroConfig
Your nest, reproducible. Layered dotfiles with Chezmoi, Hyprland, and Quickshell — copy what you need, understand why it exists.
personal DXAgentic Workstation
Four layers: machine → tools → intelligence → workspace. Health checks, doctor, and skills that survive a reinstall.
toolkitAgent Toolkit
Composable AI toolkit — 52 skills, loops, packs, and MCP templates for every assistant. One source of truth, many targets.
V ecosystemV Language
Core Team perspective: language, tooling, and ecosystem — from setup-v to the libraries that make V useful for science.
templatesCreate Awesome
Scaffold without decisions fatigue. Templates + addons that compose cleanly across Node, Python, and V.
communityCommunity & Open Source
Evidence over vanity. Contribution timelines, proof links, and pathways for beginners and maintainers.
How I work, not how I score
No skill bars. Each strength is a system: principles, tools, and proof you can verify.
Linux tooling & reproducibility
Systems that can be rebuilt from zero and audited line by line. Chezmoi, shell, and containers over hidden state.
- Dotfiles as code
- Hermetic installs & idempotent scripts
- Hyprland/Quickshell tailored, still portable
Composable CLIs
Small CLIs that compose well. Typed, tested, and documented — so they survive beyond the author.
- create-awesome-* pattern (template + addon)
- Typed boundaries (Zod at the edge)
- Helpful --help and man pages
AI agents that actually help
Agents are tools, not theatre. Skills, loops, and memory designed for solo → swarm handoffs and cost control.
- Skills > prompts (reusable, testable)
- Persona & memory for continuity
- Herdr / tmux backends with budgets
Scientific computing, precisely
Numerical code with tests, diagrams, and references — not copy-pasted formulas. Clarity over cleverness.
- VSL: structured, tested numerical methods
- VTL: tensor + autograd with graph clarity
- Docs with evidence, not claims
Open source, verified
No vanity metrics. Every claim links to a repo, commit, or org you can inspect — not a star count.
V Language — Core Team
Shipping language tooling and scientific libraries used by the V community. Contributions are public, reviewed, and linked.
Verified via github.com/vlang — member badge and authored commits
@ NaNLABS — Open-source program lead
Lead the initiative for reference repositories and boilerplates. Every public repo catalogued transparently.
20+ public repos — not vanity stars, but reusable starting points
AUR & community tooling
Packaging and maintenance where it counts — reproducible installs for the tools people actually run.
Evidence links instead of download counts
Open to collaborating
Linux tooling, CLI apps, AI agents, and science computing — if you build in those worlds, I’d love to compare notes.
Reach me
The fastest way is email or Discord. For code, open an issue on any repo — I review every one.
Collaboration CTA
Have an idea that needs a workstation, a CLI, or a scientific library? I help teams ship the first reliable version and the docs that keep it alive.
- Advisory on reproducible workstations & dotfiles
- CLI / toolkit architecture and reviews
- AI workspace & agentic harness setup
- V scientific computing — VSL / VTL
Response time: usually within 24h — or open a GH issue for async.