Execution

Engineering Team Performance

Better software starts with a better team, not a better tool.

The problem you are facing

Your sprints end with unfinished stories. Production incidents and regressions are piling up. Technical debt grows faster than it shrinks. Your best developers are starting to leave.

Your teams are experimenting with AI on their own: Copilot here, ChatGPT there, without a common method, without measuring the real gains. Legacy documentation is obsolete or non-existent. The quality vs velocity trade-off is never resolved.

You have tried adding tools, changing frameworks, hiring more. But the problem is not technical: it is the way your team specifies, develops, tests and delivers that is the real lever.

My approach: Coach-Craft

I measure before acting. My approach relies on DORA metrics, process observation, and individual interviews to build a factual diagnostic. Then I coach your team on real sprints, not in a training room.

The Coach-Craft method follows 4 steps: demonstration (I show on your real projects), guided practice (pair programming, pedagogical reviews), progressive autonomy (tech leads take over), lasting anchoring (internal champions trained to sustain). The goal is autonomy, not dependency.

Phase 1: Diagnostic (1-2 weeks)

Individual interviews, DORA metrics, process observation, ecosystem mapping (tools, workflows, AI initiatives, interdependencies). Quick win identification and sector benchmarking.

Phase 2: Embedded coaching (4-8 weeks)

Pair programming on real projects. Pedagogical code reviews. Introduction of Specification-Driven Development (SDD): specify first, generate next. CI/CD setup, quality gates, test coverage.

Phase 3: Autonomy (4 weeks)

Tech leads take over. Final KPI measurement. Complete practice transfer. Documentation of validated workflows. The team continues without me.

SDD methodology

Specification-Driven Development: AI is an executor that transforms specifications into code, not an oracle that guesses intent. Acceleration: 5-10 user stories/day, delivery cycle -30%, test coverage 80%+.

Philosophy: "Code is Cattle, Context is King"

AI can generate code faster and cheaper than any human. The value is no longer in the code itself: it is a regenerable derivative. Your true assets are your specifications and your tests. If your specs are precise and your tests complete, you can regenerate your entire codebase on demand.

Specifications

Primary asset. The "what" and the "why". Source of truth.

Tests

Primary asset. The safety net. Guarantees the "how".

Code

Regenerable derivative. An artefact, no longer an asset.

The 4 pain points every CIO knows

Every transformation engagement starts from the same ground realities, validated with dozens of CIOs in finance, insurance and public sector:

1. Understand

"What does this system actually do, in its entirety?" Not a single forgotten connection, function or stored procedure.

2. Document

"How do we capture knowledge before it retires?" No more tribal knowledge walking out the door.

3. Secure

"Where are the vulnerabilities hiding in this legacy?" Every change is a risk as long as the perimeter is not controlled.

4. Evolve

"How do we evolve without breaking everything?" The Onion Model approach: modernise from outside in, layer by layer.

Measured results from real missions

5% → 65%

test coverage

-70%

prod regressions

-96%

lead time (2d → 2h)

-73%

technical debt

Measured progression (6 months)

Test coverage 5% → 65%
Production regressions 10/month → 3/month
Delivery lead time 2 days → 2 hours
Technical debt (SonarQube) 45 days → 12 days

Results measured on a 6-month technical coaching mission (digital agency, 30 developers).

x4-x5

productivity with SDD

3-5 days

time to first sprint (vs 3-5 weeks)

80%+

target test coverage

Engagement formats

Entry point

Flash diagnostic

Ecosystem mapping, maturity benchmarks, 5-8 concrete recommendations, costed roadmap at 3, 6 and 12 months. The CTO leaves with the arguments to invest.

  • 3-5 days (including 1 day on site)
  • Deliverable: report + benchmarks + roadmap
  • CHF 5,500 - 9,000

Transformation

12-week embedded coaching

Coaching on real sprints. Pair programming, code reviews, SDD, CI/CD, quality gates. Measurable DORA metrics from month 3.

  • 12 weeks (3d/wk then 2d/wk)
  • CHF 45,000 - 55,000
  • KPIs measured: quality, velocity, efficiency

One-off

Culture workshop

3 half-days on a specific topic (code review, incident management, testing, AI for developers). Before/after survey.

  • 5-6 days preparation + facilitation
  • CHF 9,500 - 11,000

One-off

Leadership offsite

Full day facilitation for the technical leadership team. Vision, priorities, alignment, action plan.

  • 3 days (preparation + facilitation)
  • CHF 5,500

New

SDLC Assessment

Complete mapping of your development cycle (specifications, design, build, test, deploy, operate). Identification of AI accelerations per stage. Transformation roadmap with priorities and a quick win to build momentum.

  • 2-3 weeks
  • Deliverable: SDLC roadmap + priority use cases
  • CHF 15,000 - 25,000

Full journey (diagnostic + 12-week coaching + workshops + annual offsite): CHF 40,000 - 55,000 in the first year.

Who is it for?

CTO / VP Engineering

Who know their team can do better but cannot identify the concrete levers. Need a factual diagnostic and coaching on real sprints.

Teams adopting AI

Who are experimenting with Copilot or ChatGPT individually and need a common method (SDD) for AI to actually accelerate delivery.

Growing scale-ups

Whose informal processes no longer hold at scale and who need to professionalise their engineering without losing agility.

Post-acquisition / merger

Who need to align technical teams from different cultures, harmonise practices and tools.

Book a conversation

A 30-minute conversation, not a sales pitch. We will discuss your situation and I will tell you honestly whether I can help.

+41 78 222 31 59