Career transition

Data catalogs Researcher → AI Evaluation Engineer

Not generic reskilling advice, but an analysis of the distance between two specific occupations: tasks, skills, pace, money and risk.

01 · Starting distance

Transition realism index

Five factors answer a more useful question than “will it work?”: where the route is naturally strong and where proof is needed.

87%strong route

This is a strong route. The strongest support is Task similarity (96%), while the main constraint is Resilience gain (65%). The index estimates the distance between roles, not your ability.

Skill transfer89%

Task similarity96%

Entry accessibility86%

Market opportunity94%

Resilience gain65%

Starting roleData catalogs Researcher · 23%

→

Learning estimate3–6 months

→

Target roleAI Evaluation Engineer · 16%

02 · What changes in the work

Task comparison

The work shifts from Analysis and data toward Analysis and data, a 0-point change. This is the main behavioral adjustment in the move.

Data catalogs ResearcherAI Evaluation Engineer96% · profile similarity

Analysis and data

People and communication

Creation and design

Hands-on work

Control and accountability

Routine operations

Data catalogs Researcher: high-exposure tasks

AI Evaluation Engineer: high-exposure tasks

Collecting and transferring routine data34%

Preparing standard documents29%

Searching and classifying information25%

03 · Foundation and gaps

Skill-gap map

The map shows the gap between your starting point and a level you can demonstrate to an employer through work evidence—not simply “know / do not know.”

Already transferable

knowledge of the sector, terminology and typical work situations
model-quality evaluation
systems thinking
software-system understanding
debugging

Needs development

financial modelling
AI-assisted scenario analysis
valuation
return and risk analysis
a practical case for the AI Evaluation Engineer role

financial modelling

Prove it in “Working prototype: Data catalogs Researcher → AI Evaluation Engineer transition case”: include a distinct output that uses financial modelling.

3 wk

start 47%target 88%

AI-assisted scenario analysis

Prove it in “Working prototype: Data catalogs Researcher → AI Evaluation Engineer transition case”: include a distinct output that uses aI-assisted scenario analysis.

3 wk

start 46%target 77%

valuation

Prove it in “Working prototype: Data catalogs Researcher → AI Evaluation Engineer transition case”: include a distinct output that uses valuation.

4 wk

start 54%target 87%

return and risk analysis

Prove it in “Working prototype: Data catalogs Researcher → AI Evaluation Engineer transition case”: include a distinct output that uses return and risk analysis.

4 wk

start 42%target 80%

a practical case for the AI Evaluation Engineer role

Prove it in “Working prototype: Data catalogs Researcher → AI Evaluation Engineer transition case”: include a distinct output that uses a practical case for the AI Evaluation Engineer role.

4 wk

start 31%target 90%

04 · Choose a pace

Three transition scenarios

The same route affects work, money and fatigue differently. A duration without weekly effort says very little.

Keep your current job

8mo.4 h/week
139 hours total

Two short weekday sessions and one hands-on weekend block.

First applications: 6 months
Trade-off: Income is protected, but market feedback arrives later.

First apply financial modelling in the current role, then build the portfolio.

balanced default

Balanced route

5mo.7 h/week
152 hours total

Three weekly sessions: theory, practice and one end-to-end project.

First applications: 4 months
Trade-off: The pace allows market feedback without abruptly ending the current career.

After the foundation in financial modelling, move into the project and first interviews.

Accelerated entry

4mo.12 h/week
208 hours total

Four study blocks weekly, weekly practice and mentor review.

First applications: 3 months
Trade-off: The new qualification develops faster, but fatigue and a shallow portfolio are real risks.

Start applying before training ends and improve evidence every week.

05 · If the direct jump is too large

Bridge occupations

These are not mandatory stops. They matter when they provide paid experience in the new kind of work before the full move.

Data catalogs Researcher→AI Agent Supervisor→AI Evaluation Engineer

in 89%out 89%≈ 10 mo.

The AI Agent Supervisor role lets you learn part of the new task set in a more familiar context, then approach AI Evaluation Engineer with stronger evidence.

First leg →Second leg →

Data catalogs Researcher→Analytics Engineer→AI Evaluation Engineer

in 89%out 89%≈ 10 mo.

The Analytics Engineer role lets you learn part of the new task set in a more familiar context, then approach AI Evaluation Engineer with stronger evidence.

First leg →Second leg →

Data catalogs Researcher→AI Security Engineer→AI Evaluation Engineer

in 72%out 64%≈ 18 mo.

The AI Security Engineer role lets you learn part of the new task set in a more familiar context, then approach AI Evaluation Engineer with stronger evidence.

First leg →Second leg →

06 · Evidence over certificates

Portfolio project

One project cannot replace experience, but it gives an employer something concrete to discuss and shows you can finish real work.

24 hours

Working prototype: Data catalogs Researcher → AI Evaluation Engineer transition case

Take a real but anonymized situation from your current field and solve it as a AI Evaluation Engineer would. The central project task is collecting and transferring routine data.

Your advantage is domain context from Data catalogs Researcher. Make it visible: show which beginner mistakes it helps you avoid.

What the project folder should contain

A repository or interactive prototype with architecture, tests and a demo
A concise decision memo covering inputs, constraints and two rejected alternatives
A result check using measurable criteria plus one failed approach and what changed
A public 5–7-screen case study with all confidential data removed

What makes the project strong

visible use of financial modelling
a real-world problem rather than a tutorial exercise
a measurable outcome and explicit limitations
enough depth to support technical interview questions

07 · France · pay before tax

Income trajectory

In the baseline scenario, modeled income returns to the current level about 5 months after learning begins. This is a scenario model, not a pay promise.

Now€4 420

During study€4 332

First offer€4 661

+1 year€5 143

+2 years€5 840

Model horizon€7 860

Show long-term salary comparison through 2035

Data catalogs Researcher€4 420 → €5 620

AI Evaluation Engineer€5 370 → €7 860

Explore current market ↗Search target-role jobs ↗

08 · Technology horizon

How automation risk changes

The move reduces modeled automation exposure by 5 points by 2035, but the target role is not immune: its task mix also changes.

2026

23%Data catalogs Researcher16%AI Evaluation Engineer

2028

29%Data catalogs Researcher23%AI Evaluation Engineer

2030

36%Data catalogs Researcher31%AI Evaluation Engineer

2035

46%Data catalogs Researcher41%AI Evaluation Engineer

09 · An honest check

What you may not like

A good career choice is more than a list of benefits. Before studying, check whether you can live with the target role’s daily reality.

Debugging consumes real time

Much of the output is invisible until late; days include root-cause analysis, documentation and detail work.

The daily rhythm will change

The target role contains substantially more working with data and ambiguous conclusions. That can be tiring even when the occupation sounds appealing in theory.

Market pay is not first-offer pay

Even when average pay is higher, a newcomer’s first offer is usually lower. A strong project and domain experience reduce—but do not erase—the gap.

10 · Where to start

Suggested sequence

01
Review 20–30 AI Evaluation Engineer vacancies and record actual tasks, mandatory requirements and tools.
02
Define the bridge from Data catalogs Researcher: knowledge of the sector, terminology and typical work situations. Prepare two examples where this experience produced a measurable result.
03
Learn financial modelling and AI-assisted scenario analysis to the level of completing an independent practical task—not merely finishing a course.
04
Build a working prototype, publish the code in a repository, and add tests, documentation and a decision record.
05
Review 20–30 vacancies and choose only courses or certificates that repeatedly appear in employer requirements.
06
Rewrite your résumé for AI Evaluation Engineer, add the case and begin with test applications, internships, projects or adjacent tasks at your current employer.

All timelines, salaries and percentages are scenario estimates. They depend on starting skills, location, experience, weekly study time and employer requirements. Validate the route through practitioner conversations, a test project and real vacancies.