Coherent 360 Limited
AI Knowledge Graph, AI Security and Agentic Consultant
How your CV stacks up
Upload your CV to see how well it fits this job role
?%
We are building an AI-native operations management system, working with Knowledge Graphs, Agentic AI and Natural Language communications layer - built from the ground up for the realities of complex, fast-moving environments.
The Opportunity
We're looking for up to three specialists to fill the critical AI roles for our 3-month MVP sprint. This is a remote, part-time or full-time contract engagement (depending on availability) with an immediate start. You'll work directly with our CTO and founding team, mentor a remote development team, and help us deliver a demo-ready MVP in 12 weeks. For the right people, there is a real possibility of extended engagement and long-term employment with the company beyond the MVP.
Function 1 - Agentic and Architecture
The most critical hire on the project. You'll design and own the agentic architecture - the orchestration layer.
What you'll build:
- Multi-agent orchestration layer: router, specialist agents, tool registry, confidence thresholds, and human-in-the-loop confirmation gates
- ReAct loop implementation with think/tool-use/observe cycles and escalation logic
- Four V1 agents covering coordination, data intelligence, budget, and workflow
- LLM gateway (LiteLLM) for multi-provider routing and cost management
- GraphRAG retrieval pipeline combining graph traversal and vector similarity
- Eval harness from the ground up: scenario construction, LLM-as-judge grading, regression gating in CI
You must have:
- Shipped production multi-agent systems (LangGraph, LlamaIndex Workflows, or equivalent)
- LLM tool-use and tool registry design experience
- Neo4j or equivalent graph database fluency (Cypher, property-graph modelling)
- Eval harness design and ownership - not just writing tests, building the whole grading pipeline
- Async Python, idempotency patterns, agent state management
Nice to have:
- Adversarial prompt testing and injection probe suite design
- LoRA/QLoRA methodology (Year 2 scope, but good to understand the roadmap)
Reasons to use Rodeo
Start with a chat, not a search bar
Grad scheme, placement, apprenticeship? Not sure what you want yet — that's fine. Your agent talks it through with you and turns "I have no idea" into a shortlist.
It searches the market for you
Every day your agent scans the market matching roles against what actually matters to you, not just keywords on a CV.
Only hits
No noise. No "maybe this fits." Just roles with a clear explanation of why they're right — and where to focus when applying.
Function 2 - Graph Expertise
What you'll build:
- Production knowledge graph schema: entities, edges, and cardinality across complex operational domains
- Event store and propose/commit mutation pipeline with full auditability
- GraphRAG retrieval layer for AI context assembly
- pgvector integration for embedding storage alongside the graph
- Structured data parsers for entity extraction and graph population
- Idempotent processing patterns and backup verification
You must have:
- Neo4j AuraDB or equivalent graph database at production scale
- Cypher fluency - complex traversals, schema design, mutation patterns
- Event-sourcing architecture
- Postgres and pgvector
- GraphRAG pipeline experience
Nice to have:
- XML-based structured document parsing
- Overlap with the AI/agents role (graph fluency is a prerequisite for both)
Function 3 - AI Security
Handling sensitive operational data across a multi-tenant, multi-tier access model.
What you'll build:
- Seven-layer AI security model: device attestation, scoped context assembly, output filter classifier, no-exfiltration tool set, behavioural anomaly detection, sensitive-data quarantine, and audit logging
- Prompt injection defence - system prompt hardening, no-user-interpolation prompt design
- Tier-boundary penetration testing across access tiers (public → restricted)
- Output filter: entity validation, sensitive pattern detection, response volume capping
- Behavioural anomaly detection service (separate from the LLM - cannot be manipulated by the same injection)
- GDPR crypto-shredding design for immutable event stores
You must have:
- Built LLM-powered systems against the OWASP LLM Top 10
- Prompt injection attack and defence at the architecture level
- JWT scope design and API gateway enforcement
- Role-escalation and tier-boundary penetration testing
- Device attestation: Apple App Attest and Google Play Integrity
Get help with your application
Your very own career expert that helps elevate your application to the next level.
Nice to have:
- SOC 2 Type II and UK GDPR Article 30 gap analysis
- Scoped context assembly security (parameterised Cypher / graph query layer)
- Overlap with the eval/QA role (adversarial probe suites)
Eval / QA Expertise
Any of the above roles is strengthened by QA and eval experience. If you have it, tell us:
- Adversarial and injection probe suite design
- Eval scenario construction and LLM-as-judge grading pipelines
- Regression gating in CI - any agent regression blocks deployment
- Automated test coverage for multi-agent systems
Mentoring & Team Context
You'll be working with and mentoring a remote development team through the MVP sprint. This isn't a heads-down solo build - we need people who can explain decisions, set patterns, review code, and bring junior contributors up to speed on agentic systems and graph architecture. Async communication skills matter here as much as technical depth.
What We're Offering
- Fully remote - work from anywhere
- Immediate start
- 3-month initial contract - part-time or full-time depending on your availability
- Direct access to the founding team and CTO
- Real possibility of extended engagement and long-term employment for the right people
- A genuinely hard technical problem worth solving
How to Apply
Tell us:
- Which role(s) you're applying for
- A specific example of an agentic system, graph schema, or security model you've shipped - what the hard part was and how you solved it
- Your availability and preferred engagement (part-time / full-time)
- Links to any relevant work, repos, or writing
No recruiters. No agencies. Direct applications only.
We read every application and will respond to those that fit within a few days.
#AI #AgenticAI #LLM #GraphDatabase #Neo4j #LangGraph #AIEngineering #MachineLearning #RemoteWork #ImmediateStart #Contract #OperationsManagement
“It took my CV and asked me questions relevant to understanding what kind of jobs to suggest for me. Suggestions were almost perfect. Jobs were exactly what I’ve been looking for.”
Jessica, London
Skills
Location