pkm/80-sources/marche-emploi/2026-05-01-veille-emploi-devs.md
Philippe 92946d3203 feat: Add new templates for tech and employment monitoring reports
- Created a new tech and AI monitoring report template for weekly updates.
- Added an employment and developer trends report template to analyze market conditions.
- Updated CLAUDE.md to include new directories for sources and templates.
- Introduced a detailed structure for capturing key signals, OKR alignment, and actionable insights.
2026-05-11 23:34:59 +02:00

435 lines
23 KiB
Markdown
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# Veille Emploi & Devs - 2026-05-01
## Synthèse exécutive
Marché dev reprend mais **entry-level toujours cassé** : postings dev +15% depuis mi-2025, mais entry-level **45% vs 2023 peak**. AI-savvy devs gagnent **+$25K premium** ($90K$130K entry vs $65K$85K traditionnel). GitHub Copilot (4.7M users, 51% faster, $967/mois value/dev) devient standard. Juniors sans AI-skills obsoletes; seniors ayant skill orchestration agents demandés.
**Red flag :** Pipeline talent fragile. Bootcamp grads hitting mur, training programs rares. Hiring marché = "ready now" precision, pas formation. **Opportunité :** Reskill existing juniors on agents (cheaper, faster que external hiring, culture retention).
**Implication Codir :** Inverser team composition (moins juniors, plus seniors + agents). Revoir compensation policy (AI-aware premium). Investir internal academy agents (pas d'offre marché).
---
## Top 5 des Signaux Forts
| Sujet | Importance | Data clé | Implication hiring / org | Action recommandée |
|---|---:|---|---|---|
| Entry-level hiring collapse | Fort | 45% vs 2023, training programs rare | Junior pipeline = 0, reskill existing or starve | Build internal onboarding agents-first, not code-first |
| AI-savvy salary premium | Fort | $90K$130K vs $65K$85K (+$25K$45K) | Attrition risk seniors; entry tier bifurcated | Raise salary bands for AI-aware, decide positioning |
| GitHub Copilot mainstream | Fort | 4.7M users, 51% faster, 88% suggestion retention | Code quality baseline rises, hiring bar up | Assume all devs use Copilot, interview for agent skills |
| Skills shift : agents = new core | Moyen | Systems design, agent orch, cloud arch in-demand; boilerplate coding out | Senior scarcity, junior devaluation | Invest systems design training, agent patterns curriculum |
| Team composition inversion | Moyen | Optimal = 20% juniors, 50% mid, 30% senior (agents do routine) | Hiring north = seniors, south = juniors | Adjust org design, mentor ratio (seniors coach mids on agents) |
---
## Impact par Rôle
### Développeurs Juniors
**Market trend :**
Entry-level postings 45% vs 2023 peak, but market recovering (+15% overall). Bootcamp grads struggling to land first role. AI agents do routine coding (boilerplate, tests, bug fixes), which was traditional junior onboarding. **Juniors are competing with agents for the same work.**
**Salaire :**
- Traditional entry : $65K$85K
- AI-savvy entry : $90K$130K
- **Bifurcation:** strong juniors with agent knowledge price high; others struggle to find role
- Market signaling : "basic junior" skills obsolete
**Skills critiques :**
- Agent oversight (not code writing)
- Code review for agent output (validation, security)
- Prompt engineering for agents
- Systems design (learn from agents? or gaps?)
- Git workflows, testing mindset
- ~~Boilerplate coding~~ (agents handle)
- ~~Pixel-perfect frontend~~ (Figma → code agents)
**Sentiment / Attrition :**
- Bootcamp grads : demoralized ("AI took my job before I had it")
- Employed juniors : upskilled rapidly or feel obsolete
- Blind/Reddit sentiment : "junior roles don't exist anymore"
- Risk : talent exodus to other fields, or desperate acceptance of low-skill roles elsewhere
**Hiring outlook :**
- **Short-term (Q2 2026) :** Entry hiring stays suppressed. Companies hiring "ready now" senior agents-expert.
- **Medium-term (H2 2026) :** Hiring may recover IF reskilling programs prove ROI, but pure external junior hiring risky
- **Long-term :** Bifurcated market. Top juniors (AI-aware) absorb entry roles; others (non-AI) pushed to other sectors
**Action recommandée :**
1. **If you're hiring :** Don't recruit external juniors; recruit seniors or reskill internals
2. **If you have juniors :** Mandatory agents curriculum (week 1 onboarding), pivot to agent supervision + systems design
3. **If you're junior :** Bootcamp alone insufficient. Seek agent-first programs or internal ramps.
---
### Développeurs Mid-Level
**Market trend :**
Mid-level hiring steady, recovering. Mid-levels are the "sweet spot" : agents handle routine, mids orchestrate + deliver features. No crisis, but upskilling pressure.
**Salaire :**
- Traditional mid : $100K$140K
- Senior-track mid : $120K$160K
- **Inflation minimal** ; market tight but no runaway comp
**Skills critiques :**
- Agent orchestration (multi-step workflows, debugging agent failures)
- Systems design (midway to senior)
- Cloud architecture
- Team coordination (agents + humans)
- Incident response (agent failures in prod)
- Security review (agent-generated code)
**Sentiment / Attrition :**
- Generally positive : work becomes more strategic (less grind, more design)
- But : orchestration complexity new overhead
- Risk of feeling "supervising bots, not building" (psychological impact)
**Hiring outlook :**
- **Strong demand.** Mid-levels are the lever for team scaling with agents.
- Easier to hire external mids than juniors or seniors
**Action recommandée :**
1. Invest in agent orchestration training (LangGraph, multi-agent patterns)
2. Create "agent orchestration" role clarity (what does this dev own?)
3. Monitor sentiment (supervision fatigue = real risk)
---
### Développeurs Seniors / Tech Leads
**Market trend :**
Senior demand **high** (agents need human architects). But supply constrained. Salary inflation significant. Competition fierce for senior agents-expert talent.
**Salaire :**
- Traditional senior : $150K$200K+
- Senior + agent expertise : $180K$250K+
- **Premium :** +$30K$50K for agents-expert architects
- C-level engineering roles pulling talent (CTO, VP Eng)
**Skills critiques :**
- Multi-agent system design (complex orchestration)
- Governance & safety (agents at scale = risk)
- Cost optimization (token budgets, compute efficiency)
- Security (agent code review, guardrails)
- Team architecture (how many agents per senior? supervision model?)
- Hiring strategy (how to attract senior talent in agent era?)
**Sentiment / Attrition :**
- **Opportunity mindset :** agents = new frontier, exciting
- But : **validation overhead risk** (19% some seniors due to code review friction)
- **Headhunting pressure :** AI-native startups aggressively recruiting senior agents architects
- **Burnout risk :** If poorly supported, seniors feel "firefighting agent failures" instead of architecture
**Hiring outlook :**
- **Very high demand,** but **supply scarce.** External hiring = expensive ($200K+ total comp).
- Internal promotion from mid-level = primary source
**Action recommandée :**
1. **Retention urgency :** Seniors are flight risk if compensation/mission unclear. Lock in equity, mission clarity.
2. **Invest in systems design training** for high-performing mids (succession pipeline)
3. **Competitive compensation :** Monitor market ($180K$250K for agents expert). Underpay = attrition.
4. **Define role clearly :** "Senior architect" in agent era ≠ old "senior dev". Clarity reduces frustration.
---
### Chefs de Projets / Product Managers
**Market trend :**
PM roles stable, but requirements changing. PM must understand agent capabilities/limitations to scope realistic roadmaps. Delivery speed up, estimation difficulty up.
**Impact :**
- Feature delivery : **70% timeline** (weeks → days for complex features)
- Quality variance : agents great routine, risky novel logic
- Scope creep risk : faster delivery tempts over-ambitious roadmaps
- Team dynamics : agents + humans = coordination complexity (more sync points, not fewer)
**Skills demanded :**
- Agent capability scoping (what can/can't agents do?)
- Risk assessment (agent failures in prod = user-facing)
- Roadmap estimation with agents (less predictable than humans)
- Team orchestration (agents + humans working together)
**Action recommandée :**
1. Upskill on agent patterns (LangGraph, multi-step workflows)
2. Create estimation framework : "agent tasks" vs "human tasks" (different velocity)
3. Plan more, scope aggressive (deliver faster, learn faster)
---
## Marché Emploi : Données Quantifiées
### Job Postings & Hiring Trends
| Métrique | Trend | Data source |
|---|---|---|
| Job postings dev (total) | +15% since mid-2025 (recovery) | ZipRecruiter / Indeed |
| Entry-level postings | 45% vs 2023 peak (suppressed) | ZipRecruiter |
| Senior/lead postings | +25% vs 2023 (high demand) | LinkedIn / ZipRecruiter |
| Hiring pace | Slowing slightly (precision hiring, not volume) | Blind, company reports |
| **Market size 20262029** | $24B (2024) → $61B (2029), +20% CAGR | Morgan Stanley / McKinsey |
**Interpretation :** Market growing fast (+20%/an), but **entry-level suppressed** (agents displace routine work). Senior talent increasingly scarce. Hiring = precision (revenue-driving roles), not broad junior pipeline.
---
### Salaires & Compensation
| Profil | Salaire USD | Premium AI | Trend | Source |
|---|---:|---:|---|---|
| Entry-level (non-AI) | $65K$85K | | Flat or ↓ (supply up, entry jobs ↓) | Levels.fyi, ZipRecruiter |
| Entry-level (AI-aware) | $90K$130K | +$25K$45K | ↑ (demand > supply) | Levels.fyi |
| Mid-level (architecture) | $120K$160K | +$15K$25K | ↑ (agent expertise) | Levels.fyi |
| Senior (agent orchestration) | $180K$250K+ | +$30K$50K | ↑↑ (scarce) | Levels.fyi, LinkedIn |
| **GitHub Copilot users (avg)** | **$95.3K** (Feb 2026) | **~$25K** (estimated) | **Proxy for AI-aware salary** | ZipRecruiter |
**Interpretation :** Clear bifurcation. AI-savvy devs earn $25K$50K more. Traditional junior track = declining comp. Senior agent-expert = highest demand + highest pay.
---
### Skills In-Demand & Trending
| Skill | Trend | Why (AI agent shift?) | Hiring demand |
|---|---|---|---|
| Python | ↑ | Agents + AI frameworks | Very high |
| TypeScript / Node.js | → | Still web/API standard | Very high |
| Systems design | ↑↑ | Agents need architecture, less junior coding | Very high |
| Agent orchestration | ↑↑↑ (new) | Multi-agent patterns (LangGraph, etc.) | Emerging high |
| Cloud architecture | ↑ | Infrastructure for agents, multi-cloud | High |
| Cybersecurity / secure coding | ↑ | Agent-generated code = audit risk | High |
| DevOps / SRE | ↑ | Agent monitoring ≠ monolith monitoring | High |
| Rust (systems) | ↑ | Infrastructure, performance-critical | Moderate |
| ~~Boilerplate coding~~ | ↓↓ | Agents handle | Low |
| ~~Pixel-perfect frontend~~ | ↓ | Figma → code agents | Moderate (UI still human) |
| ~~Routine bug fixing~~ | ↓ | Agents handle | Low |
**Interpretation :** Architect + systems design skills premium. Routine coding skills depreciate. Agents create demand for "how to supervise agents" knowledge (new skillset).
---
### Talent Sentiment & Attrition
| Signal | Data | Implication |
|---|---|---|
| Junior morale (Blind, Reddit) | "entry-level jobs don't exist", bootcamp grads struggling | Attrition to other fields? Reskilling inertia? |
| Senior burnout (agent oversight?) | "validating agent code is tedious", mixed sentiment on agents | Validation overhead real. Tool + process improvements needed. |
| Bootcamp grad employment | Placement rates down 3040% vs 2023 | Pipeline fragile. Bootcamps need to teach agents-first. |
| Mid-level satisfaction | Generally positive (less grind, more strategy) | Retention stable if compensation fair. |
| Avg tenure senior roles | Slight ↓ (headhunting pressure from AI startups) | Senior attrition risk if underpaid or misaligned. |
**Interpretation :** Juniors demoralized (pipeline crisis). Seniors at flight risk (attractive market). Mids stable. Overall : talent market bifurcated, not unified.
---
## Implications pour Stratégie Hiring
### Entry-Level Hiring Strategy
**Current state :**
- Hiring external juniors = high-risk, low-reward (no role fit, training overhead)
- Bootcamp pipeline broken (45% postings)
- Training programs rare (companies hire "ready now")
**Challenge :**
- Agents do routine coding (traditional junior work)
- Juniors can't learn coding-by-doing anymore (no grind tasks)
- Market expects "junior able to supervise agents day 1" = unrealistic
**Recommendation :**
1. **Stop external junior hiring** (at least through 2026). Use freed budget for senior hires + reskilling.
2. **Build internal junior program** :
- Recruit **interns or bootcamp-adjacent talent** (cheaper, lower bar)
- Curriculum = agents-first (week 1 : Copilot, LangGraph, prompt engineering)
- Not "write CRUD API" but "supervise agent writing CRUD API"
- 12-week ramp (vs 6-month traditional), then productive
3. **Or : partner bootcamp** (negotiate curriculum focus on agents)
**Timeline :** Q2 2026 design, Q3 2026 launch first cohort.
---
### Mid-Level Hiring Strategy
**Current state :**
- External mid hiring = viable (supply better than junior, demand high)
- Mid-level = "sweet spot" for agent era (agents + human orchestration)
**Demand :**
- Agent orchestration experience (multi-step workflows, error handling)
- Systems design (mid-level arch thinking)
- Team leadership (supervising agents + juniors)
**Recommendation :**
1. **Hire for "agent orchestration" explicitly** (not generic "mid-level dev")
2. **Interview questions should include :**
- "How do you debug a multi-agent workflow?"
- "Design a system where agents handle X, humans handle Y"
- "How do you validate agent-generated code?"
3. **Look for candidates with Claude / ChatGPT / Copilot experience** (signal of agent-aware thinking)
**Salary :** $120K$160K depending on agent expertise.
---
### Senior Hiring Strategy
**Current state :**
- External senior hiring = expensive but essential (supply < demand)
- Agent-architect expertise = rare, premium-priced
**Demand :**
- Multi-agent system design (orchestration, safety, cost)
- Technical strategy (agents vs traditional code trade-offs)
- Team architecture (how many agents per human?)
- Hiring & mentoring (building agent-era teams)
**Recommendation :**
1. **Hire for "agent system architect" roles** (new job title emerging)
2. **Compensation :** $180K$250K+ (market competitive)
3. **Look for :**
- Claude / GPT API experience (not just Copilot)
- Proven agent deployment in prod (rare but exists)
- Leadership (shipping teams, not solo IC)
4. **Poach from :** AI-native startups, OpenAI/Anthropic partners, research labs
5. **Retention :** Lock in equity, mission clarity, compensation review (agents-expert premium = real)
---
### Compensation Policy Update
**Current salary bands :**
(Assume your bands before reading this)
**Market shift :**
- AI-savvy premium = +$25K$50K real, not perception
- Senior agent architects = highest demand, highest pay
- Entry-level bifurcation (AI-aware vs traditional)
**Recommendation :**
1. **Raise entry-level ceiling** (if you hire AI-aware juniors) : $85K $100K+ (to attract bootcamp grads willing to ramp)
2. **Raise mid-level range** : Add "agent orchestration expert" track above mid baseline (+$15K$25K)
3. **Raise senior ceiling** : $200K $230K+ for agent architects (external benchmark = $250K)
4. **Create explicit "AI-aware" tier** in your levels (communicate market reality)
**Timing :** Q2 2026 audit, Q3 2026 implement (before H2 hiring sprint).
---
## Implications pour Structure Équipes
### Current Team Composition (estimated for 50-dev shop)
| Level | Headcount | % Team | Role |
|---|---:|---:|---|
| Junior | 25 | 50% | Routine coding, learning |
| Mid | 15 | 30% | Feature delivery, mentoring |
| Senior | 10 | 20% | Architecture, hiring, strategy |
---
### Recommended Composition (Agent Era, same 50-dev shop)
| Level | Headcount | % Team | Role | Rationale |
|---|---:|---:|---|---|
| Junior | 10 | 20% | Agent supervision, code review, learning systems design | Agents do routine; juniors oversee + learn architecture. Fewer, but higher-skilled. |
| Mid | 25 | 50% | Feature delivery, agent orchestration, team coordination | Core delivery layer. Agents + mids = main velocity. |
| Senior | 15 | 30% | Architecture, agent system design, hiring, mentoring | Design multi-agent workflows. Build team. Grow mids. |
**Implications :**
- Hire differently : fewer juniors, more seniors + mids
- Mentor differently : seniors coach mids on "how to orchestrate agents", not "how to code"
- Onboard differently : juniors start with agents, not blank slate coding
**Timeline :** Transition over 612 months (organic promotion of strong mids, external senior hiring, junior hiring freeze).
---
### Team Structure Evolution Needed
**Org Design :**
- Create "agent orchestration lead" role per squad (senior mid or staff engineer)
- Pair each agent system with 23 human supervisors (prevent 1-person bottleneck)
- Create "agent reliability team" (debugging, monitoring, incident response)
**Career Ladder :**
- **New titles :** "Senior Agent Architect" (vs "Senior Dev")
- **Mid path :** "Agent Orchestration Specialist" (track record of multi-agent systems)
- **Junior path :** "Agent Supervisor" (first role, learning systems design)
**Mentoring :**
- Senior = coaches mids on "how to scope agent tasks, design agent workflows"
- Mid = coaches juniors on "how to validate agent output, review code for logic errors"
- Not "here's how to write perfect code", but "here's how to work with AI"
---
## Red Flags & Opportunities
### Red Flags
| Flag | Evidence | Risk | Mitigation |
|---|---|---|---|
| Junior pipeline evaporating | Entry-level 45% vs 2023, bootcamps struggling | Can't grow talent internally, external hiring breaks down | Build internal reskilling track (week 1 agents curriculum), partner bootcamp on agent-first programs |
| Salary compression | AI-aware +$25K premium emerging, entry-level bifurcating | Attrition seniors? Or entry tier entirely obsolete? | Audit compensation, decide positioning (hire AI-aware premium or traditional track?) |
| Agent orchestration skill gap | No one trained, market has few experts | Hiring externals risky; internal upskilling slow | Invest 812 weeks in agent patterns workshop (LangGraph, prompt eng, multi-step workflows) |
| Validation overhead (seniors 19% productivity) | Some seniors slowed by agent code review | Burnout risk; seniors lose edge if too much validation | Tool investment (automated checkers for agent code), process (what should humans review vs skip?) |
| Headhunting pressure on seniors | AI startups aggressively recruiting | Attrition of best architects | Lock in equity, mission clarity, compensation survey (stay competitive $250K band) |
### Opportunities
| Opportunity | Why now | Expected benefit | Effort |
|---|---|---|---|
| Reskill existing juniors on agents | Entry market broken, but internal talent exists, agents-first curriculum clear | Cheaper hiring ($15K onboarding vs $20K recruiting), culture retention, agents expertise | **Moyen** (48 weeks training, ongoing curriculum) |
| Acquire senior architect talent | High demand for agent design, budget freed from skipping external junior hiring | Build moat vs competitors on agent orchestration, attract future talent (reputation) | **Fort** (recruiting, offer $200K+, onboarding) |
| Invest in internal academy (agents) | No market training exists; bootcamps struggling to adapt | Become known for agent talent, attract top developers, competitive advantage | **Fort** (6 months build, ongoing delivery) |
| Adjust org design early (fewer juniors, more mids) | Trend clear, but competitors slow to adapt | Talent utilization optimization, more strategic work per person, culture shift to architecture | **Moyen** (org change is friction, but doable over 6 months) |
---
## Questions à Creuser
1. **Hiring strategy :** How do you build a junior pipeline when entry-level jobs vanish? Is internal reskilling realistic, or is bootcamp partnership necessary?
2. **Compensation :** Do you raise salary bands to match AI-aware market, or accept attrition of strong juniors to other fields?
3. **Team math :** If agents +3050% productivity, and juniors can't learn coding anymore (no grind tasks), what's the optimal team ratio (juniors:mids:seniors)? Is 20% juniors realistic?
4. **Retention :** Which of your seniors are at flight risk (headhunting pressure)? How do you keep them (mission, comp, equity)?
5. **Culture :** Does "your senior became an architect of agent swarms" feel like promotion or job evolution? Is team excited or anxious?
6. **International :** Is this trend US-only (Levels.fyi, Blind = US-centric data) or happening in France / EU? Salary trends different?
7. **Competitors :** Are your competitors moving on this (reskilling, hiring seniors)? Or are you ahead of the curve?
---
## À Intégrer dans PKM
| Note | Action | Priorité |
|---|---|---|
| 20-areas/pro/management/hiring-strategy-2026 | **Créer :** Framework for agent-era hiring (junior internal ramp vs external senior hunt) | ⭐⭐⭐ |
| 20-areas/pro/management/compensation-2026 | **Audit :** Current salary bands vs market. Decision : AI-aware premium or traditional track? | ⭐⭐⭐ |
| 20-areas/pro/management/junior-pipeline | **Create :** Internal bootcamp curriculum (agents-first, 12 weeks, hire interns/early career) | ⭐⭐⭐ |
| 20-areas/pro/cto/team-structure-agent-era | **Document :** Ideal org composition (20% juniors, 50% mids, 30% seniors). Timeline for transition. | ⭐⭐ |
| 20-areas/pro/cto/senior-retention | **Plan :** Compensation + mission + equity review for agents-expert seniors (flight risk). | ⭐⭐ |
| 30-resources/management/reskilling-agents-curriculum | **Create :** 812 week curriculum for juniors : LangGraph, prompt eng, multi-step agents, validation patterns. | ⭐⭐ |
---
## Sources Consultées
| Source | Éditeur | Date | Type |
|---|---|---:|---|
| [2026 Agentic Coding Trends Report](https://resources.anthropic.com/hubfs/2026%20Agentic%20Coding%20Trends%20Report.pdf) | Anthropic | May 2026 | Report |
| [Agentic Coding in 2026: AI's Impact on Software Development](https://www.timesofai.com/industry-insights/agentic-coding-in-software-development/) | Times of AI | 2026 | Article |
| [The demise of software engineering jobs has been greatly exaggerated](https://www.cnn.com/2026/04/08/tech/ai-software-developer-jobs) | CNN Business | April 2026 | News |
| [GitHub Copilot Statistics 2026](https://www.getpanto.ai/blog/github-copilot-statistics) | GetPanto | 2026 | Data |
| [Developer Job Market Recovery 2026: Data Analysis and Trends](https://pooya.blog/blog/software-dev-job-market-recovery-2026/) | Pooya Golchian | 2026 | Analysis |
| [Salary: Github Copilot (February, 2026) United States](https://www.ziprecruiter.com/Salaries/Github-Copilot-Salary) | ZipRecruiter | Feb 2026 | Data |
| [AI Impact on the Job Market in 2026: What the Data Shows](https://www.secondtalent.com/resources/ai-impact-job-market-2026/) | Second Talent | 2026 | Report |
| [Tech Job Market 2026: Trends, Skills, and Opportunities](https://legacy.anitab.org/blog/resources/tech-job-market-2026/) | AnitaB | 2026 | Article |
---
## Actions pour le Mois Prochain
- [ ] Audit compensation vs market (Levels.fyi, LinkedIn)
- [ ] Sketch internal junior bootcamp (agents-first curriculum)
- [ ] Interview 35 senior candidates (understand agent-architect expectations)
- [ ] Sync with HR/People (hiring strategy shift, junior hiring freeze implications)
- [ ] Decide : internal ramp vs external hiring for juniors