Dragonfly Cryptocurrency Recruitment Current Situation Survey: Compliance Roles +340%, Data Science +74%, Cryptocurrency Enters the "On-Demand Hiring" Era

Author: Zackary Skelly (Dragonfly Talent Lead)

Translation: Deep Tide TechFlow

Deep Tide Guide: Dragonfly releases the 2026 Talent Insights report for the crypto industry, revealing a fundamental shift in recruitment logic. In 2025, the industry net laid off 472 people, but compliance roles surged by 340%, and data science roles grew by 74%. The most critical change is: candidates are no longer impulsive in a bull market; they want clear value explanations and certainty. If you can’t clearly explain “why this position is important,” conversion rates will plummet.

1/

We have entered the first quarter of 2026, and recruitment in the crypto space is completely different from any previous cycle.

We just released the latest “Talent Insights” report, detailing how we got here and what it means for founders and talent teams.

2/

TL;DR

2025 did not kill crypto recruiting but made it more mature.

Companies are no longer hiring based on price but based on actual needs.

This shift has become the new benchmark entering 2026.

3/

This year is clearly divided into two halves.

The first half of 2025 (25H1) was turbulent, with macro shocks quickly reversing the optimistic sentiment toward crypto.

Job removals surged in March (750), most of the losses concentrated in the first half.

Throughout the year, about 3,700 new positions were added, and about 4,100 removed, net decrease of -472.

4/

The second half (H2) brought discipline and recovery.

The overall trend line of positions in H2 is similar to 2024, just at a lower level.

July reset, August bottomed out, September reopened, and Q4 stabilized.

The more intense reset in spring was the main reason why 2025 overall was below 2024.

5/

In our 25H1 report, we made some predictions. Now, let’s score them:

✓ Q3 rebound in late Q3 (positions opened in September +26%), Q4 slowdown, compliance hiring started early

✗ Underestimated the divergence in traffic and applications; overestimated the resilience of legal roles relative to compliance roles

6/

From 25H1 to H2, the real shift isn’t how many people companies hired, but what roles they hired. Core priorities, first to win expansion rights.

→ Engineering: -12%, still the anchor → Marketing: -27% → Design: -33% → Customer Support: -35% → Sales & BD: -16% → Legal: -41% → Compliance: +340%

7/

Data science was the most obvious winner of the year, with a +74% YoY increase. (Thanks to AI? ))

8/

Interesting changes also occurred on the candidate side.

Traffic remained stable in H2, but applications dropped about 26%.

People are still browsing, but no longer submitting applications easily.

9/

In early cycles, market excitement drove most of the hiring: rising salaries, influx of applications.

This mechanism is failing.

Stronger months can still boost views, but attention-to-conversion rates are no longer as high as before.

10/

Why? Partly because candidates have become more cautious.

They scrutinize company durability, ownership clarity, team quality, and technical credibility more strictly: open-source proof, product depth, hardcore issues, GTM roadmap.

Vague category narratives no longer work.

11/

Areas of concentrated belief: infrastructure, DeFi, L1 and L2 remain core, but DeFi interest narrows to stablecoins, payments, and RWA.

Fintech-related and institutional use cases gained significant attention. AI remains a key interest point.

12/

Stage preferences also tell an interesting story.

Seed and Series A stages remain most attractive to candidates, with high demand for founder roles and first hires. However, larger, more mature companies still attract interest.

13/

The main factors causing candidate attrition are not salary, stage, or size, but ambiguity.

If you can’t clearly explain why the company is important, what specific scope they will have, and why the opportunity is sustainable, conversion rates will drop sharply.

14/

Geographically, remote remains the norm, but the most active recruiting teams are concentrated in New York, with a preference for face-to-face work.

Talent remains globalized, but New York + Bay Area still dominate. Europe is the largest non-US hub.

(Note: Specific geographic recruiting = smaller TAM, longer hiring cycles. ))

15/

Another factor shaping the current landscape is: recruiting is increasingly concentrated on later-stage teams, especially in areas candidates find most interesting.

We expect that the remaining hiring in 2026 will be driven more by acquisitions, transformations, and integrations rather than pure growth.

16/

So, what should founders do?

Recruit based on milestones—product launches, revenue, partnerships, regulatory progress—rather than market cycles or calendar plans.

Companies that recruited well in H2 2025 clearly articulated and stuck to the reasons each role existed.

17/

Knowing that teams differ, plan personnel sequences carefully:

→ Prioritize core builders (engineering, security, data/protocol) → Explore BD fit → Flexibility in product focus (earlier for consumer, leaner for infrastructure) → Compliance, finance, risk → Marketing/support scaled after leverage appears

18/

Maintain open pipelines for scarce talent.

Supply of engineering, AI/ML, and security roles is severely constrained; you can’t restart from zero each cycle. Even if specific needs are closed, relationships should stay warm.

19/

Recognize that role sales methods have changed.

Candidates want clear career paths, ownership within 30–60 days, and transparent upside mechanisms.

You must sell differentiation. You’re not just selling your category but why you’ll win and what specific roles they can play.

20/

You also need a genuine AI story. Not “We are an AI company.”

Candidates want to know:

→ How AI is used internally → How it changes the product → Whether it creates real advantages

Vague answers will lose talent.

21/

Specific advice for talent teams:

Put your strongest people at the front of the process (first impressions matter), keep interview loops tight, and provide clear feedback.

22/

An open question: AI makes 2026 harder to predict.

People can do more with fewer staff. Better tools enable some to start their own ventures. Some may go directly into AI work.

Meanwhile, higher individual productivity means faster scaling, and crypto positioning is broader than ever.

23/

Our current view on AI’s impact: before clear AI × Crypto use cases solidify, slowdown signals outweigh acceleration signals.

📎 Further reading: The Agentic Economy Will Be Massive, Agentic Commerce Won’t

24/

Our baseline expectation for 2026: flat to modest growth, led by engineering, AI/data, and security. Integration will continue.

Whether bull, baseline, or bear market, this is a year focused on quality building.

25/

Teams that win talent will be those with the most credible stories, not the loudest voices.

Discipline in execution, durable business models, and good explanations of both are essential.