Ibraheem Abdul-Malik

Engineering the systems that let AI agents do real work

Senior AI engineer designing and shipping autonomous agent platforms, clinical AI systems, and intelligent developer tools. I build the infrastructure that lets AI agents plan, code, review, and ship software with human oversight.

What I Build

Multi-Agent Orchestration

Built a desktop platform that coordinates autonomous AI coding agents in organizational hierarchies. Agents plan, implement, and review code in isolated git worktrees with budget gating, session resumption, and governance-aware delegation.

Agent OrchestrationGit WorktreesBudget GatingDesktop-First

AI Developer Tools & SDKs

Created plugin architectures, MCP server integrations, and agent skill frameworks (21+ behavioral guardrails) that let teams extend AI capabilities into project management, source control, and custom internal tools.

MCP ServersPlugin SDKsAgent SkillsSDK Design

Clinical AI & Healthcare

Applied machine learning to clinical workflows at scale, building decision support systems and real-time data pipelines that serve thousands of healthcare providers surfacing insights from complex medical data.

Clinical AIData PipelinesHIPAAProvider Tools

Infrastructure & Architecture

Engineered scalable cloud infrastructure for AI workloads including real-time inference pipelines, vector databases, and CI/CD systems. Designed desktop-first architectures with local-first data and no cloud dependency for core operations.

Cloud InfraIaCLocal-First DataReal-time Systems

Select Highlights

  • Shipped an AI agent platform where a CEO agent delegates tasks to engineer and reviewer agents, each working in isolated git worktrees with automatic PR creation and budget enforcement
  • Built clinical AI tools used by thousands of healthcare providers to surface insights from complex medical data, with HIPAA-compliant pipelines processing millions of records
  • Created bidirectional integrations between agents and project management systems, with agent tools for searching, creating, and syncing issues plus status sync, comment bridging, and webhook handlers
  • Designed an agent skill framework with 21 behavioral guardrails covering build, verify, review, ship, and operate phases of the development lifecycle
  • Architected a multi-model adapter interface enabling agents to swap between frontier LLM providers without vendor lock-in