An engineering team working together at their desks

AI-Native Engineer Bootcamp

[01]DASA Bootcamps · AI-Native Engineer

Use AI agents safely from commit to production

A hands-on program for development and operations engineers. Participants configure agentic harnesses, verify generated changes, and define controls for production use.

  • 4 weekly sessions
  • 14 hours
  • Live, online
  • Up to 20 participants

Facilitated by Dr. Rik Farenhorst and Samuel Kaluvuri

[02]DASA Bootcamps · The shift

The constraint now sits in verification and safe operation

AI has sharply reduced the cost of producing code and operational artifacts. Delivery now slows at review, integration, observability, and accountability. The program teaches engineers to find the bottleneck and reduce the work waiting there, rather than generating more code.

  • Generated code and automation reach review faster than teams can verify and release them.
  • Agents enter repositories without shared instructions, bounded permissions, or reliable verification.
  • When generated changes reach production, ownership and blast radius remain unclear.

[03]DASA Bootcamps · Outcomes

By the end of the bootcamp, you'll be able to

Diagnose where AI changes delivery cost and where work now waits.

Configure an agentic harness and verify the changes it produces.

Make undocumented systems safer to change and identify missing telemetry.

Define provenance, accountability, and autonomy controls for production use.

Built for

  • Software developers
  • Operations engineers
  • Platform engineers
  • SREs
  • QA engineers
  • Security engineers
  • Software architects

[04]DASA Bootcamps · Curriculum

Bootcamp curriculum

Four weekly live sessions of 3.5 hours. Groups of four or five work hands-on in a shared course repository, without needing access to a client production system.

Module 1: Cheap to write, expensive to trust

Understand how AI changes the economics of engineering work and moves the delivery constraint downstream.

  • Compare the falling cost of generation with the remaining cost of verification.
  • Identify where faster output creates new queues and risk across delivery.

Module 2: Where your changes wait

Map how a change moves from commit to production and find the queue that limits throughput.

  • Use batch size, wait time, and work in progress to diagnose delivery flow.
  • Produce one agreed statement naming the dominant constraint.

Module 3: Finding your real numbers

Measure delivery and operational performance using data from tools the team already owns.

  • Derive lead time and deployment frequency from existing delivery data.
  • Connect recovery and reliability measures to the same delivery flow.

Module 4: Setting up the agent to do real work

Configure an agentic harness that gives the agent useful context and clear operating boundaries.

  • Define conventions, permissions, verification commands, and what done means.
  • Drive one specification-first change through the configured harness.

Module 5: Reviewing more than you can read

Adapt review and quality gates to the volume and confidence of AI-generated work.

  • Separate plausible output from evidence grounded in tests and system behavior.
  • Design verification gates that still bind when generation accelerates.

Module 6: The systems nobody wants to touch

Make legacy code and undocumented automation safe enough to understand and modify.

  • Use characterization tests to pin current behavior before changing it.
  • Use AI to identify assumptions, then verify them against tests and observed behavior.

Module 7: Seeing into what you run

Build the telemetry required before AI-assisted operational analysis can be trusted.

  • Identify the signals missing from the system and prioritize the instrumentation backlog.
  • Connect code changes with the behavior of the service in production.

Module 8: Who owns code no human wrote

Make provenance and accountability explicit for generated changes.

  • Record how work was generated, verified, approved, and promoted.
  • Focus governance on the AI risks engineering teams face now.

Module 9: Letting it touch production

Set autonomy levels for agents that can interact with running systems.

  • Distinguish between actions that read, propose, and act.
  • Limit how much an agent can affect reliability, security, or cloud cost.

Module 10: What changes on Monday

Turn the program into concrete changes to engineering work and team responsibilities.

  • Identify the platform and team structure changes required for safe adoption.
  • Set owned, measurable commitments for the next 90 days.

[05]DASA Bootcamps · What you take away

Four outputs your team can use immediately

Each group produces four connected artifacts. Each becomes an input for the next and remains useful after the program.

Constraint diagnosis

A map of how work moves from commit to production, ending with an agreed statement of the dominant queue.

Harness configuration

Committed instructions covering conventions, permitted scope, verification commands, and completion criteria.

Characterization tests and telemetry gaps

Tests that pin current behavior, together with a prioritized list of missing operational signals.

Provenance and autonomy policy

Rules for acceptable use, provenance, and autonomy, supported by 90-day commitments with named owners and measures.

[06]DASA Bootcamps · Certification

The path to certification

The optional exam tests judgment against the ten program outcomes, independent of any specific AI product. It has 40 applied judgment questions, is web-proctored, and is scheduled after the program.

01

Complete the program

Finish the full 14-hour curriculum and the four connected build blocks.

02

Prepare and schedule

Use the exam preparation in the final session, then schedule the web-proctored exam.

03

Earn the credential

Score 60% or above on the 40-question exam to earn the DASA AI-Native Engineer certification.

[07]DASA Bootcamps · How to join

Two ways to join

Open cohort

Join a public cohort

Live, instructor-led online, 14 hours over 4 weeks, with engineers from other organizations. Leave with four linked artifacts you can apply at work. Contact us for dates and pricing.

Register interest

In-house

Private team cohort

Bring up to 20 development and operations practitioners together. Keep the certified core while adapting examples to your stack and context, and turn the course artifacts into a shared 90-day implementation plan.

Book a conversation

See all bootcamps