Tech Radar

A personal, opinionated radar of the technologies, techniques, platforms and tools I actually practice — by Jean-François Beaulieu, Director of Innovation and Software Architect. It makes good on my mission: to hold an active technology watch and a living radar that separates durable signal from noise.

It is not an objective market ranking. Every blip is a judgment grounded in practice, at a point in time, and meant to be revised. Disagreement is welcome — that’s what an opinionated radar is for.

How to read it

The radar follows the ThoughtWorks-style convention: four quadrants and four rings.

Quadrants say what kind of thing it is:

• Techniques — architecture practices, patterns, applied-AI methods, governance and compliance, agentic engineering.
• Platforms — clouds, runtimes, self-hosting platforms, LLM platforms and models, execution infrastructure.
• Tools — developer tooling, AI tools, observability, the architect’s productivity stack.
• Languages & Frameworks — languages, application frameworks, AI / agentic frameworks.

Rings say what I recommend doing about it. They are a gradient of decision maturity, not popularity:

• Adopt — proven by real use; my default choice in the right context.
• Trial — used on a real project; promising enough to try where the risk is absorbable.
• Assess — actively watching and exploring; no project commitment yet.
• Hold — don’t start anything new with this: disappointed after use, obsolete, or hype without durable substance. A negative signal grounded in experience — never merely “too new” (the untried belongs in Assess).

What earns a blip

A technology enters only if it passes three cumulative tests:

1. Evidence — I have an opinion grounded in practice (tried it, watched it fail, or watched it hold), not from a blog post I read. The bar rises with the ring.
2. Specificity — I can name it precisely. If I can only name a category, I don’t have a blip.
3. Durability — the mission filter: a signal of substance that will change practice, not a hype cycle forgotten in two quarters. Declining a fashionable technology is a stronger signal of judgment than adopting it.

Blips are deliberately scarce: a lean, defensible radar beats a bloated one. Each blip carries a short justification including its tradeoff — what it costs, not only what it gives.

Cadence

Target quarterly, with a non-negotiable floor of two editions per year. Each edition is a dated snapshot; at every edition every existing blip is reviewed (promoted, demoted, or retired), not just appended. A stale radar is worse than no radar.

Trajectory

Positions move: Assess → Trial on the first real project use, Trial → Adopt when not using it would become unreasonable. The Adopt → Hold flow — disavowing an earlier choice — is the richest signal. Each blip’s trajectory stays visible on its own page; a blip that is stable and no longer relevant leaves the radar (an active view, not a graveyard).

---

Read the essays behind the radar →