Skip to content
View johnnichev's full-sized avatar
🌌
Building new things!
🌌
Building new things!
47 followers · 55 following

Achievements

Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: Pull Sharkx3Achievement: QuickdrawAchievement: Arctic Code Vault Contributor

Achievements

Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: Pull Sharkx3Achievement: QuickdrawAchievement: Arctic Code Vault Contributor

Block or report johnnichev

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
johnnichev/README.md

John Nichev

Senior ML Engineer with 10+ years building scalable systems and ML infrastructure across startups and enterprise.

I build production-grade tools that make AI work reliably in real codebases, not demos.

Currently shipping

The nv: skills collection. Production AI methodology, distilled into one-command installable Claude Code skills.

Live experience: skills.nichevlabs.com · production case studies, the research library, and the full synthesis.

Skill What it does
nv:context Context engineering for AI coding agents. 200+ research sources, 8 distilled laws, validated on 3 production repos. Case studies · Research · Synthesis
nv:design Vibe Design methodology. Uses source code as design reference instead of prompts. Built skills.nichevlabs.com end to end.
nv:dev Plan, test, debug. The development workflow loop for AI coding agents.
nv:ops Guardrails, evaluation, multi-agent orchestration. 
npx skills add johnnichev/nv-context -g -y

Flagship open source

Selectools

Production-ready Python framework for AI agents with built-in guardrails, audit logging, and cost tracking. Powers the NichevLabs orchestration layer behind multiple SaaS products.

  • 5 LLM providers · 146 models
  • 4,612 tests · unit, integration, regression, E2E with real APIs
  • Tool calling, prompt injection defense, execution traces, sessions, memory
  • Hybrid BM25 + vector search, semantic chunking, cross-encoder reranking
  • Property-based testing (Hypothesis), thread-safety smoke suite, production simulations

What I work on

  • AI agent infrastructure · tool calling, guardrails, prompt injection defense, execution traces, multi-provider orchestration
  • RAG and search · hybrid BM25 + vector search, semantic chunking, cross-encoder reranking
  • Full-stack engineering · React, Next.js, Node.js, TypeScript, Python, FastAPI
  • Data and ML pipelines · PyTorch, recommendation engines, analytics pipelines
  • Cloud and infrastructure · AWS, GCP, Docker, Kubernetes

Reach me

For engineers who ship.

Pinned Loading

  1. selectools selectools Public

    Production-ready Python framework for AI agents with built-in guardrails, audit logging, cost tracking, and hybrid RAG. Supports OpenAI, Anthropic, Gemini, Ollama. By NichevLabs.

    Python 9

  2. nv-context nv-context Public

    nv:context > Context engineering for engineers who ship. Auto-detects, analyzes, generates configs, hooks, session management, and token budgets. 200+ research sources.

    Shell 3

  3. nv-design nv-design Public

    nv:design — Professional web design with AI. Design system extraction, landing pages, site modernization, image/video prompts. Vibe Design methodology.

    1

  4. nv-ops nv-ops Public

    nv:ops — Guardrails, evaluation, multi-agent orchestration. Agent infrastructure: safety, measurement, scaling for production AI agents.

  5. nv-dev nv-dev Public

    nv:dev — Plan, test, debug. The development workflow loop for AI coding agents. Spec-driven development, TDD with property/mutation testing, systematic debugging.

  6. nv-context-landing nv-context-landing Public

    Landing page for nv:context — context engineering for engineers who ship. Built in one afternoon with nv:design.

    HTML