Senior AI Agent Engineer (LangGraph / MCP)
As Senior AI Agent Engineer at Esaholic, you will own the core architecture for enterprise autonomous multi-agent swarms. You will build deterministic state graphs using LangGraph, implement PostgreSQL checkpointers for zero-data-loss execution, develop custom Model Context Protocol (MCP) servers, and enforce human-in-the-loop governance.
What You Will Accomplish in Your First 90 Days
We do not hire engineers to sit in endless planning meetings. From week one, you will own critical components of our agent orchestration infrastructure.
Graph State Debugging & Tool Isolation
- Audit our core LangGraph state schema and ship performance fixes to `PostgresSaver` checkpointer queries.
- Build and deploy your first custom Model Context Protocol (MCP) server over SSE transport.
- Pair with Umar Abbas to review production agent trace benchmarks.
Multi-Agent Swarm Engineering
- Architect a complex multi-agent state graph with cyclic error-recovery edges and conditional routing.
- Implement `interrupt_before` human-in-the-loop approval webhooks for high-risk transactional nodes.
- Benchmark agent execution latency under 500+ parallel thread connections.
Enterprise Client Deployment
- Deploy production agent clusters for our enterprise banking or fintech clients.
- Establish automated regression testing suites for agent tool calling accuracy.
- Publish an internal technical post-mortem guide on state graph failure modes.
Systems & Technologies You Will Own
- LangGraph Orchestration Core: Python state graphs, custom `TypedDict` state schemas, and `Annotated` reducer logic.
- Durable Checkpointer Subsystem: `PostgresSaver` and `RedisSaver` thread state persistence engines.
- MCP Tool Server Ecosystem: JSON-RPC 2.0 servers connecting PostgreSQL, REST APIs, and legacy systems to LLMs.
- Human-in-the-Loop Gateways: Asynchronous approval webhooks and authorization control flows.
- Languages & Runtimes: Python 3.12+, FastAPI, Pydantic v2, AsyncIO.
- Database Engines: PostgreSQL 16 (`pgvector`), Redis 7, psycopg-pool.
- Inference Backends: vLLM GPU serving clusters, OpenAI API endpoints, Anthropic Claude.
- DevOps & Deployment: Docker, Kubernetes, Helm, GitHub Actions CI/CD.
What We Do NOT Expect from You
We do not demand "5 years of LangGraph experience" - LangGraph was released in 2024. We care about your fundamental understanding of state machines, Python concurrency, and distributed systems.
We will never ask you to invert a binary tree on a whiteboard under pressure. Our interview focuses 100% on practical system design, real Python code walkthroughs, and state graph design.
We do not run brutal 24/7 on-call rotations. Production deployments are protected by automated health checks, self-healing container pods, and strict SLO SLAs.
You will not be asked to polish CSS or build complex frontend frameworks. Your domain is backend AI agent architecture, state graphs, and API integrations.
How We Interview Candidates
Our hiring process is fast, respectful of your time, and completed within 14 calendar days from initial call to offer decision.
Senior AI Agent Engineer Hiring Process Roadmap
Phase Delivery RoadmapIntro Screen
Conversation with engineering team discussing your background, career goals, and technical philosophy.
- ✓ Mutual Alignment Check
- ✓ Role Detail Review
System Design & Code
Practical state graph design discussion. We present a real agent problem and review architecture together.
- ✓ State Machine Architecture
- ✓ MCP Tool Integration
Architect Deep-Dive
Technical conversation with Umar Abbas (Founder & Principal AI Architect) covering concurrency, DB checkpointers, and production trade-offs.
- ✓ Deep Architecture Sync
- ✓ Team Culture Check
Offer & Decision
Formal offer letter detailing base salary, equity options, equipment budget, and start date options.
- ✓ Signed Offer Letter
- ✓ Onboarding Plan
Text alternative for screen readers & search engines
- Phase 1: Intro Screen (30 Mins (Call)) - Conversation with engineering team discussing your background, career goals, and technical philosophy. Key deliverables: Mutual Alignment Check, Role Detail Review.
- Phase 2: System Design & Code (60 Mins (Video)) - Practical state graph design discussion. We present a real agent problem and review architecture together. Key deliverables: State Machine Architecture, MCP Tool Integration.
- Phase 3: Architect Deep-Dive (45 Mins (Video)) - Technical conversation with Umar Abbas (Founder & Principal AI Architect) covering concurrency, DB checkpointers, and production trade-offs. Key deliverables: Deep Architecture Sync, Team Culture Check.
- Phase 4: Offer & Decision (48 Hours) - Formal offer letter detailing base salary, equity options, equipment budget, and start date options. Key deliverables: Signed Offer Letter, Onboarding Plan.
Compensation & Benefits Package
- Base Salary: £105,000 to £135,000 GBP per annum (based on technical experience).
- Equity Options: Direct equity option grant in Esaholic.
- Remote Work Setup: £2,500 workstation budget (MacBook Pro M3 Max / dual 4K displays).
- Learning & Conference Budget: £1,500 annual budget for AI conferences and research papers.
- Time Off: 28 business days paid annual leave + UK bank holidays.
Location & Work Policy
This role is Remote-First for candidates residing in the United Kingdom or European time zones (GMT ± 3 hours). Candidates residing in London have optional access to our hybrid office space.
How to Apply
Send an email to our engineering team with your CV and links to your GitHub profile, technical blog, or published code repositories:
Subject Line: Application: Senior AI Agent Engineer - [Your Name]
Include: CV PDF, GitHub link, and a 2-sentence note on your most interesting Python or agent project.