Associate Forward Deployed Engineer

Cognitio Labs; Singapore; Hybrid; SGD 40,000 - 70,000; Full time

Job summary

Cognitio Labs in Singapore offers an Associate FDE role where you embed with a client, solve real operating problems, and participate in a three-month paid apprenticeship on live client work. A full-time offer follows for those who carry the work, with opportunities to grow within the Singapore centre.

Eligible candidates peer into applied AI, will train with senior engineers, and join a program that starts with a three-month paid apprenticeship and progresses to a full-time position based on

Qualifications

• No formal degree requirements; open to polytechnic/ITE/university candidates.

Responsibilities

• Embed deeply with one client to understand operations and data flows.
• Run the deployed agent in daily work and report back improvements.
• Collaborate with senior engineers to expand capabilities and impact.

Skills

Learning ability

Problem solving

Client liaison

Job description

1.3521° N / 103.8198° E
Built for Singapore. Deployed from Singapore.

We train Forward Deployed Engineers. Then we hire them.

From knowledge to intelligence. CognitioLabs takes applied AI research through to produce bespoke agents that run across multiple industries and sectors. Forward Deployed Engineers are the people who put them there, and who stay until the team uses them without being asked.

Course → 3 months paid, in the field → Full-time role

Applications open Training centre / Singapore Scroll

01

P / 01

It ends in a job

Awareness, exposure and a short attachment are already well served here. This is the rung after: the course, then three months of paid apprenticeship on live client work, then a full-time offer to the ones who can carry it.

02

P / 02

Trained, not imported

Forward Deployed Engineering is taught, not hired away. We run the curriculum, practice apprenticeships, and teach on the job on real cases.

03

P / 03

Field over classroom

The hard part was never the model. It is sitting with an operating team until you understand how the work really happens. Every module ends in reviewed work on a live problem.

Recruiting across every path

Junior Colleges out of scope, age and National Service timing

ROUTE / 01

/// Why FDE exists

Capable models do not become useful systems on their own. FDEs identify the right workflow, calibrate model limits and build the context, controls and operating loop required for reliable use.

01

Market signal / 01

Capability without calibration

Leaders treat AI as a universal answer while operating teams reduce it to a chat tool. Neither view reveals the real capability boundary.

02

Market signal / 02

Use cases without diagnosis

Teams ask for an AI feature before identifying the workflow, bottleneck and business outcome that actually matter.

03

Market signal / 03

Agents without operations

A promising demo is not a solution. It still needs context, integration, evaluation, monitoring and updates as the business changes.

/// Your role

What you do as an
Associate FDE.

An Associate FDE embeds deeply with one client. Solve one operating problem, run the system, then use the context you gained to find the next. Each loop goes deeper and advances the enterprise's AI maturity.

Loop 01 / 04 · Learning users / systems / decisions

01

Embed in the business

Work inside one client environment long enough to understand its people, data, constraints and decision patterns.

02

Find the next problem

Turn operational friction into the next valuable AI use case, not a disconnected feature request.

03

Build into the workflow

Connect models, data, tools, interfaces, evaluations and controls until the system runs in daily work.

04

Operate and deepen

Maintain the agent, measure real use and turn what you learn into the next, deeper problem.

Loop
+1

What we actually check

A polished portfolio mostly records who had access, encouragement and spare time, so screening on one would just re-sort people by where they started.

• You can pick up something unfamiliar quickly
• You can stay with a problem you cannot yet solve
• You want to be in the room with real users

Deliberately not required

Shipment exceptions arrive through email, chat and spreadsheets. Your job is to map the triage workflow and turn scattered updates into a controlled operating loop.

01 / Before

Chat updates

Spreadsheet status

02 / The system

Ingest + context

Structured triage

Human approval

Historical cases become an eval set. Every recommendation keeps human approval and an audit trail.

03 / Evidence

Decision accuracy

Associate FDE owns

Senior FDE signs off

/// Selection

Show how you build.
Show how you recover.

AI tools are allowed. We care whether you can explain, change and test the work. Candidate exercises are never used in production.

What we assess

Engineering + debugging Problem framing Evaluation Communication Learning speed

/// Partners

We do not build
this alone.

A training centre only works when the schools, the employers and the people funding it are pointing the same way. We run the engineering. The rest of it depends on who stands next to us.

Schools and educators

Send us the ones who would never find this

Nomination beats open application. An open call only reaches students who already knew the opportunity existed, which is the wrong filter. Put a name forward and we handle the curriculum, the supervision and the placement.

Employers

You get an engineer embedded in a real operating problem, working under a named senior reviewer who signs off before anything ships. They get the one thing no course can manufacture, which is the actual room.

Foundations and community partners

Fund capability, not headcount

Human capital work is hard to measure, so we measure the part that is not ambiguous. Every cohort reports named outcomes: apprenticeship starts, conversions to full-time, starting salary, retention at twelve months. Published whether the numbers are good or not.

The intake is deliberately small. A handful of people placed properly is worth more than a large programme with soft results.

2 to 3

Paid apprenticeships by Q1 2027

2+

Full-time conversions by Mar 2027

100%

Builds reviewed by a senior engineer

If your organisation works on human capital, youth pathways or getting Singaporeans into the work that is coming, there is a version of this we should build together.

Office and client-site work, not remote-only

Travel

Occasional APAC travel depending on the engagement

Work eligibility

/// FAQ + application

Practical details on location, deployment scope and how selection work is handled.

No. A science, mathematics, engineering or physics foundation is enough, and we recruit across the polytechnics, ITE and the universities equally. What we are checking is whether you learn quickly and stay with a problem you cannot yet solve.

Yes, and that is deliberate. A polished portfolio mostly records who had access, encouragement and spare time. We would rather teach you something unfamiliar in the room and watch how you use it, which is what the selection process is built to do.

Yes. Three months of paid work on live client deployments, not an unpaid placement, with a senior engineer reviewing every build before it reaches a client. At the end of the three months the apprentices who can carry the work are offered a full-time role.

You will be given problems you cannot solve alone, on purpose, with someone senior alongside you. Everyone who completes the course keeps their work, a written reference and the professional contacts they made.

Yes. Use the tools you would use to build well. You must be able to explain, modify and test anything you submit.

The centre exists to build capability inside Singapore, so the roles are based here and we recruit locally first. Any work pass requirement is considered case by case.

Applications open

The course, then three months of paid apprenticeship on live client work, then a full-time offer to the ones who can carry it.