Services
A structured way to modernise production.
For teams running real systems who need to move faster without breaking what works. The engagement runs as a sequence: Blueprint → Build → Calibrate.
Most of the work sits across cloud architecture, data pipelines, and platform reliability—tightening observability, improving cost optimisation, and adding automation where it reduces toil. Where it fits, that includes AI integration: connecting models to production systems with the same constraints and rollback controls as everything else.
Packages
Clear scope, published pricing
Three engagement types, each with defined outputs and ownership transfer. Pick the shape that fits your constraint.
What you get: A system map, documented baselines, and a sequenced plan with guardrails—before any build work starts.
- System map
- Baseline metrics
- Risk register + guardrails
- Sequenced plan
- Handover notes
Not a fit for greenfield MVPs with no production signals yet.
What you get: Production changes with baselines, rollback paths, and runbooks—shipped in controlled slices.
- Baselines first
- Controlled rollouts
- Rollback paths
- Runbooks
- Measurable change
Not a fit if constraints and scope aren't clear—start with Blueprint.
What you get: SLOs, cost budgets, regression controls, and an operating rhythm your team owns.
- SLOs + alerts
- Performance + cost budgets
- Regression controls
- Operating rhythm
- Exit to internal ownership
Not a fit for systems still mid-build or without stable baselines.
Final scope set after a short intake call. Price varies by environments, compliance, access constraints, and on-call needs.
BlackLake scales through method, not noise—maintaining the same standards as scope increases: baselines before change, controlled rollouts, operational clarity, and systems designed to endure.
Blueprint
Fixed-scope discovery
A paid, time-boxed assessment that makes constraints explicit and produces a plan you can run.
- Current-state system map (boundaries + contracts)
- Baseline metrics (latency, cost, failure modes)
- Risk register with guardrails + rollback plan
- Sequenced delivery plan (milestones + decision points)
- Handover notes: ownership, operating model, next steps
Build
Delivery with controlled change
Implementation work that ships safely: baselines first, then changes in slices with clear rollback paths.
- Delivery in controlled slices with measurable baselines
- Interfaces hardened: boundaries, contracts, and integration
- Rollout controls: canaries, feature flags, rollback paths
- Workstreams across systems, intelligence, and operator tooling
- Handover that includes runbooks and ownership boundaries
Calibrate
Reliability, optimisation, ownership
Keep production stable while tightening performance and cost. The goal is an operating rhythm you own.
- Reliability work: SLOs, alert quality, and failure-mode containment
- Performance + cost budgets with ongoing measurement
- Operational tightening: runbooks, incident patterns, and drills
- Regression controls: checks that prevent silent drift
- Exit criteria: internal ownership with clear operating rhythm
Capabilities
Focused delivery across cloud, AI, and full-stack.
Each capability is built around measurable baselines, guardrails, and explicit trade-offs that keep production operable.
Cloud Architecture & Infrastructure
Change is blocked by brittle infrastructure, unclear ownership, and rising operational load.
Stabilise interfaces, deployment, and run-time control so change ships without regressions or surprise spend.
- Baselines + telemetry
- Guardrails + rollout
- SLOs + alerting
- Runbooks + handover
Applied AI & Machine Learning
Models ship without evaluation, drift detection, or a plan for failure in production.
Apply AI where it measurably reduces human load, with explicit constraints, baselines, and safe fallbacks.
- Evaluation + drift checks
- Fallbacks + guardrails
- Auditable behaviour
- Serving budgets
Full-Stack Development
Operators work around the tooling, and every release adds new failure modes.
Build operator-facing tools with performance baselines, clear ownership, and safe releases.
- Operator-first workflows
- Performance baselines
- Observable critical paths
- Safe releases
FAQ
Questions we get asked
Straight answers on scope, access, pricing, and how the work runs.
Why publish price ranges instead of fixed quotes?
Fixed quotes before understanding constraints are guesses. The ranges reflect real variance: a straightforward Blueprint in a well-documented system costs less than one spanning five legacy services with no tests. After the intake call, you get an exact number.
Do we have to start with a Blueprint?
Not always. If you already have a clear system map, documented constraints, and a sequenced plan, we can move straight to Build. Most teams find the Blueprint valuable because it surfaces risks and scope gaps before delivery starts—but it's not mandatory.
What access do you need to start?
For Blueprint: read access to repos, architecture docs, and 2–3 hours with the team who owns the system. For Build: a dev environment, CI access, and a path to staging. We never require production credentials upfront—access is scoped to what the current phase needs.
How do you control risk in production?
Every change ships with a rollback path. We use feature flags, canary deploys, and staged rollouts. Baselines are measured before changes land so regressions show up in metrics, not incident calls. Guardrails and circuit breakers are part of the delivery, not afterthoughts.
What does a Build sprint look like?
Two to four weeks of focused delivery. Each sprint starts with a baseline and ends with measurable change in production. You get daily async updates, a mid-sprint check-in, and a demo at the end. Handover includes runbooks and ownership notes—not just merged code.
What does Calibrate include, and when do you exit?
Calibrate covers reliability work (SLOs, alert quality), cost and performance budgets, and operational tightening (runbooks, incident patterns). Exit happens when your team can run the operating rhythm without us—usually 3–6 months, depending on complexity and internal capacity.
Can we pause or change scope mid-engagement?
Yes. Blueprint is fixed-scope, but Build sprints can be paused or redirected between sprints. Calibrate is month-to-month. If priorities shift, we adjust—no long-term lock-in.
How to start
Begin with constraints, not commitments
Share context and what must change. If it’s a fit, the next step is a Blueprint: a short engagement that produces a scoped plan and guardrails.