Full-Stack Software Engineer | AI Agents • Developer Tools • SaaS

I build and ship production software across the stack, from backend systems and infrastructure to frontend applications and AI-powered developer tools.

My recent work focuses on agentic AI, local LLMs, developer tooling, and production SaaS, with an emphasis on building practical systems that automate real work.

• Languages: TypeScript • JavaScript • Python • Java • Ruby • SQL
• Frontend: React • Next.js • Tailwind CSS
• Backend: Node.js • Ruby on Rails • Spring Boot • PostgreSQL • Redis
• AI: LLMs • AI Agents • MCP • Tool Calling • Ollama • Local Inference
• Infrastructure: Docker • Linux • CI/CD • AWS • Cloudflare • Vercel

Distributed software engineering harness for autonomous coding agents

Building an agent orchestration platform that turns feature specifications into verified software changes. ForgeLoop coordinates specialized agents across isolated execution environments and uses builds, tests, containers, browser automation, and independent review to close the loop between generated code and working software.

• Multi-agent orchestration with dependency-aware task graphs and parallel execution
• Isolated Git worktrees and self-hosted runners for agent execution
• Closed-loop verification with automated failure detection, repair, and re-verification
• MCP-based tooling, scoped agent permissions, and targeted repository context
• Provider-agnostic architecture for hosted APIs, BYOK workflows, and local models
• Evidence-backed runs tracking tests, tool calls, artifacts, acceptance criteria, and execution metrics
• Java/Spring Boot control plane with React, GraphQL, PostgreSQL, Docker, and Playwright
• Independent Ticketly Spring Boot + React SaaS repository for realistic end-to-end agent testing

Evidence-first AI research and writing platform

Built a research system where AI agents can plan research, discover and use MCP tools, evaluate sources, and synthesize findings while maintaining an evidence trail behind generated claims.

• Agentic research workflows with planning, tool discovery, execution, evaluation, and synthesis
• MCP integration for connecting agents to external research tools and data sources
• Evidence trails linking claims to sources, excerpts, tool calls, confidence, and limitations
• Workspace-based research sessions with saved runs, sources, and revision history
• APA and MLA paper generation with citations and export support
• Next.js frontend, Node.js API, PostgreSQL, Firebase authentication, Stripe billing, and Claude API

Local-first, provider-agnostic coding agent platform

Built an agent runtime that works across desktop, VS Code, CLI/TUI, and Neovim interfaces.

• Local and cloud LLM support through Ollama, LM Studio, llama.cpp, and compatible APIs
• MCP, tool calling, structured outputs, Git operations, terminal execution, and browser preview
• Shared configuration and agent behavior across multiple clients
• Designed around local-first and BYOK workflows

AI-powered resume tailoring platform

Full-stack AI SaaS for tailoring resumes to specific job opportunities while preserving the user's actual experience.

• Python backend with a TypeScript frontend
• LLM-powered resume analysis and tailoring
• Firebase authentication with Google and email/password
• BYOK support, usage credits, and Stripe subscriptions
• Independently deployed frontend and backend infrastructure

B2B lead generation and automation platform

Built and operate a production SaaS for discovering, processing, and managing business leads.

• Distributed browser automation and data-processing workflows
• Optimized collection jobs for 30–50% faster execution and 50–70% lower memory usage
• Stripe billing, Firebase authentication, HubSpot and Zapier integrations
• Chrome extension, exports, email workflows, and production infrastructure

I like owning problems end to end.

That can mean designing a database schema, building an API, implementing the frontend, integrating an LLM, debugging a production issue, configuring infrastructure, or figuring out why something is using twice as much memory as it should.

I'm particularly interested in the intersection of software engineering and AI, especially systems where models can use tools, interact with real environments, and perform useful work rather than simply generate text.

• 🌐 Portfolio: https://portfolio.hookerhillstudios.com
• 💼 LinkedIn: https://www.linkedin.com/in/jaredhooker

Currently open to Full-Stack Software Engineer and AI Engineer opportunities.