How we work

Engineers scope it. Build it. Prove it.

The engineers who scope your solution are the ones who build it. When it's ready, you choose: your team runs it, or we do.

Engagement models

Two ways to work with us.

The difference is who operates your solution after go-live. Choose a project-based engagement with a handover to your team, or ongoing managed services from Nebula Logix.

Option 1Build it. Own it. Run it.

Fixed-Scope Integration

We integrate and deliver your solution within your existing environment. Once the agreed deliverables are accepted, your team takes responsibility for day-to-day operation and maintenance.

  • Deployed in your cloud environment using agreed tools and architecture.
  • Handover of source code, infrastructure-as-code, documentation and runbooks.
  • Knowledge transfer so your team can operate and maintain the solution.
  • Defined scope, deliverables and acceptance criteria.

Pricing: A fixed project fee based on the agreed scope of work.

Ongoing support, enhancements and third-party usage costs are scoped separately unless included in the agreement.

Option 2Build it. We operate it.

Ongoing Managed Integration

We integrate your solution across the agreed environments and provide ongoing operational support, monitoring and maintenance under a defined service agreement.

  • Runs in your cloud environment, as agreed.
  • Monitoring and security tooling based on the solution's architecture and requirements.
  • Ongoing maintenance, troubleshooting and agreed upgrades.
  • Defined service coverage, response targets and escalation procedures.

Pricing: A recurring monthly management fee.

Monitoring and security tooling we operate is included or itemized, as agreed. Cloud usage is billed to you directly by your cloud provider.

Pricing after scoping call.

Delivery

Scope → Ground → Orchestrate → Prove → Run

Typically seven weeks from agreed scope to a go-live decision. Run applies to retainer engagements only.

Delivery phases
01 / 05Phase
00

Scope

Week 0

A 30-minute call with an engineer on the integration problem. We agree the scope and how we'll measure it.

You get
  • Agreed scope
  • Pricing after the scoping call
  • Agreed success measures
01

Ground

Weeks 1–2

We map your systems, data flows and failure points, and agree the baseline.

You get
  • System and data-flow map
  • Access agreed with your IT team
  • Baseline: errors, manual hours, latency
02

Orchestrate

Weeks 3–5

We build the integrations and workflows around your real process in our development and validation environments, with human or LLM-as-a-judge review wherever AI acts.

You get
  • Integrations built and validated in our dev environment
  • Human or LLM-judge review on every AI step
  • Logging, tracing and alerts wired in
03

Prove

Weeks 6–7

We deploy to your cloud environment and test there with your real data, in UAT / staging, then clear the production bar before go-live.

You get
  • Deployed to your UAT / staging environment
  • Tested on your real data, in your cloud
  • Evaluation report, security evidence pack and go-live decision
04

Run

Retainer only

Retainer engagements only. Ongoing LLMOps: we monitor every agent call and trace in Langfuse, check that responses stay grounded in your sources, and tune prompts, retrieval and models as your data changes. On fixed-scope work, your team runs it from handover.

You get
  • Agent calls and traces monitored in Langfuse, with alerts
  • Grounding and quality checks on live responses
  • Tuning, regression tests and model upgrades
Go-live gate

Grounded · Evaluated · Least privilege · Reviewed · Security-reviewed · Observable · Handed over

Security

Your account. Your keys. Your logs.

Your cloud account
Private network
Serverless workflowFunctions and orchestration
AI, in your regionManaged AI modelsYour account, your region
Human or LLM-judge reviewYour reviewers approve
Your data and connectorsSearch index and integrations in your account
Key managementYour keys
Audit logsEvery action logged
Owned by your organization
Engineer access · revocable · loggedNebula teamScoped access, authorized by you
  • Runs serverless in your cloud account, on AWS or Azure, with AI models in your region.
  • Encrypted with keys you own and manage.
  • Our access, and every AI agent's access, is an identity you issue (Okta or AWS IAM) and can revoke at any time; every action appears in your own audit logs.
  • Your content is never used to train models. No lock-in.
  • We don't resell cloud. You pay your cloud provider directly, at their price.

We don't certify anyone's compliance. Your CMMC status, SSP and ATO stay yours; we build systems that fit inside them and provide the evidence to show how.

Questions

About engagements

What's the difference between the two options?

Who runs it after go-live. With Fixed-Scope Integration we build in your environment and hand over; your team operates it. With Ongoing Managed Integration we operate, monitor and maintain it for a monthly fee.

Do we have to use your tools?

No. Fixed-scope work runs in your environment with your tools. In managed engagements we can operate monitoring and security tooling for you, or use what you already run. You pay your cloud provider directly.

How is it priced?

Pricing follows the scoping call: a fixed price for the agreed scope, or a monthly fee for managed integration. Never hourly. Cloud usage is billed to you by your provider.

Do you resell cloud services?

No. You pay your cloud provider directly, at their price.

Have an integration problem?

30 minutes with an engineer, no slides. We agree the scope on the call and send pricing after.