Vaibhav Mule's avatar

Forward Deployed Engineer

Vaibhav Mule

vaibhavmuleplank-fde-trainingJoined November 2026

Skills

  • AI workflow
  • Models and prompts
  • Agents and retrieval
  • Product and interfaces
  • Agentic RAG
  • Browser automation
  • Claude Code
  • MCP
  • Multi-agent systems
  • OpenAI SDK
  • Prompt engineering
  • Tool use
  • Agent design
  • Agent frameworks
  • Agent handoffs
  • Agent patterns
  • AI agents
  • AI dev environment
  • AI SDLC
  • AI web scraping
  • Anthropic SDK
  • Chat app architecture
  • Claude Agent SDK
  • Claude Code skills
  • CLAUDE.md
  • Component catalogs
  • Computer use
  • Context management
All 49 skills
  • Conversation memory
  • Cost tracking
  • Custom MCP servers
  • Embeddings
  • Error recovery
  • Feature delivery
  • Generative UI
  • Git worktrees
  • GitHub workflow
  • Hooks and slash commands
  • LLM cost control
  • Long-term memory
  • Multi-turn tool loops
  • Next.js
  • Parallel agents
  • Progressive rendering
  • Prompt chaining
  • Retrieval
  • Review-driven development
  • Spec writing
  • Storage patterns
  • Streaming
  • Streaming UI
  • Subagents
  • Vercel

Timeline

  1. Day 15 · Week 3: Agents, retrieval, browser

    Shipped

    Turned the analyst into an orchestrator delegating to knowledge and web specialist agents, with a shared findings list, timestamped per-agent logs, per-agent cost tracking and scripted tests.

    • Multi-agent systems
    • Agentic RAG
    • Browser automation
    • Error recovery
    • Cost tracking

    The research analyst: orchestration, retrieval and the browser

  2. Day 14 · Week 3: Agents, retrieval, browser

    Shipped

    Added a schema-driven AI scraper that follows next links, and a screenshot-driven Chrome browser agent tested against a legacy-portal fixture, with CLIs.

    • AI web scraping
    • Browser automation
    • Computer use

    AI scraping, browser use and desktop use

  3. Day 13 · Week 3: Agents, retrieval, browser

    Shipped

    Grew the knowledge base to 52 documents, added chunking with tsvector/pgvector search, find/ls/cat/grep filesystem tools, and a top-5 versus agent comparison script.

    • Agentic RAG
    • Embeddings
    • Retrieval

    Modern RAG: agentic retrieval

  4. Day 12 · Week 3: Agents, retrieval, browser

    Shipped

    Built a 4-agent reply pipeline (researcher, writer, editor, orchestrator) on the Claude Agent SDK with guardrail and trace hooks, plus a triage intake API.

    • Multi-agent systems
    • Agent handoffs
    • Claude Agent SDK
    • Agent frameworks

    Multi-agent systems and frameworks

  5. Day 11 · Week 3: Agents, retrieval, browser

    Shipped

    Built a from-scratch agent loop (runAgent) with step, dollar and time caps, an investigate agent with policy search and findings, and a CLI /investigate command.

    • AI agents
    • Agent patterns
    • Agent design

    AI agent fundamentals

  6. Day 10 · Week 2: SDK, chat, generative UI, memory

    Shipped

    Added per-visitor long-term memory in Postgres with fact extraction, a 24,000-token context budget with summarization, and a settings page to forget memories.

    • Context management
    • Conversation memory
    • Long-term memory
    • Storage patterns

    Context and memory

  7. Day 9 · Week 2: SDK, chat, generative UI, memory

    Shipped

    Added json-render with a 22-component Zod catalog, a streaming /api/generate route, and a /dashboards generator with three example scenarios.

    • Generative UI
    • Component catalogs
    • Progressive rendering

    Generative UI

  8. Day 8 · Week 2: SDK, chat, generative UI, memory

    Shipped

    Added a streaming /api/chat route via the AI SDK with Day 7 tools, a chat page with Markdown, tool-call states, localStorage conversations, and admin assistant.

    • Next.js
    • Streaming UI
    • Chat app architecture
    • Tool use

    Full-stack AI chat app

  9. Day 7 · Week 2: SDK, chat, generative UI, memory

    Shipped

    Added a tool registry (calculator, read_url, save_note, web_search), tool loops for Anthropic and OpenAI-compatible providers, and a /research chain with citations.

    • Tool use
    • Multi-turn tool loops
    • Prompt chaining
    • OpenAI SDK

    Tool use and chaining calls

  10. Day 6 · Week 2: SDK, chat, generative UI, memory

    Shipped

    Added Anthropic and OpenAI streaming providers with pricing, a WorkBot system prompt, a session with model switching and cost, and a CLI chat.

    • Anthropic SDK
    • OpenAI SDK
    • Streaming
    • Prompt engineering
    • LLM cost control

    Anatomy of an AI app

  11. Day 5 · Week 1: AI-powered development workflow

    Shipped

    Moved requests into Postgres/PGlite, added PostHog intake-funnel events, and built tested database and PostHog MCP servers wired via .mcp.json.

    • MCP
    • Custom MCP servers
    • Feature delivery

    MCP servers and complete feature development

  12. Day 4 · Week 1: AI-powered development workflow

    Shipped

    Shipped parallel features WB-12 to WB-15: request detail page, status updates, team/status filters, Markdown policy library, plus folder-level CLAUDE.md files and a policies API.

    • Git worktrees
    • Parallel agents
    • CLAUDE.md
    • Prompt engineering

    Multi-project and multi-feature workflow

  13. Day 3 · Week 1: AI-powered development workflow

    Shipped

    Created a ticketed backlog (WB-5 to WB-11) and shipped request model, intake form, requests list, validated API, admin navigation and settings move, each with tests.

    • AI SDLC
    • MCP
    • Spec writing
    • Review-driven development

    The AI SDLC: feature development workflow

  14. Day 2 · Week 1: AI-powered development workflow

    Shipped

    Added three Claude Code skills (component, api-route, pre-deploy), three reviewer subagents with a pr-review skill, and a formatting hook in settings.json.

    • Claude Code
    • Claude Code skills
    • Subagents
    • Hooks and slash commands

    Claude Code: skills and subagents

  15. Day 1 · Week 1: AI-powered development workflow

    Shipped

    Shipped eleven small Next.js changes: /api/health route, settings deployment page, shadcn badge, WorkBot landing page, 404, favicon, robots.txt, metadata, footer.

    • Claude Code
    • AI dev environment
    • GitHub workflow
    • Vercel

    Dev environment and Claude Code fundamentals