The short version
- Anton Alexander questioned the skills assumptions behind Trinidad and Tobago's AI data-centre ambitions in an August interview. His personal assessment deserves attention, but it is not a national workforce survey.
- The government's July estimate of more than 5,000 jobs covered three proposed projects, including steel. It was conditional, not a count of AI vacancies or local people hired.
- Data-centre work includes several occupations. Match a specific role to its actual entry requirements before paying for a course.
- Ask employers and training providers for a dated plan showing the work, supervised experience, recruitment stages and costs. A course title alone cannot establish that connection.
If you are an electrician, a software developer or a student considering further study, “AI jobs” is too broad a description to plan around. You need to know what the work involves, who might employ you and what you would have to learn. Those questions matter even when a proposed facility is still being negotiated.
What the engineer's warning establishes
In a Trinidad Express interview published on 18 August 2026, Anton Alexander warned that large AI facilities require scarce specialist expertise. The newspaper identified him as a Trinidad-born senior generative AI specialist at Amazon Web Services and explicitly said he was speaking personally, independently of AWS. He supported pursuing the sector while questioning assumptions about the available skills and resulting benefits.
AWS's own author biography describes Alexander's work on large training and inference workloads, CUDA programming and Kubernetes. That helps readers understand his technical background. It does not turn his interview into an AWS assessment of Trinidad and Tobago, or establish how many qualified people live here.
The distinction matters. A knowledgeable person's warning can identify questions worth investigating. To size a local skills shortage, an employer would still need a list of required roles, evidence about existing workers and a recruitment plan. This review did not find those answers in the interview and did not conduct that workforce assessment.
Get the jobs figure into the right category
The government statement published by TTT News on 11 July described three frameworks signed the previous day. They concerned EY's proposed collaboration with third parties on a 300 MW data centre, Hummingbird's proposed 150 MW facility with possible expansion to 500 MW, and Pinnacle's steel project. The combined potential investment exceeded US$5 billion, with more than 5,000 new jobs expected if the projects advanced.
Those totals covered all three proposals. They do not establish the number of permanent data-centre posts, their locations or the share open to people with any particular qualification. Hummingbird's early-2028 operating target was also subject to due diligence and the project advancing. Treat the July statement as dated evidence of the plans announced then.
Before using a jobs estimate to choose training, ask for a breakdown by project and occupation. Separate construction work from continuing operations and distinguish direct employment from work with suppliers. A family deciding whether to fund further study needs a different level of detail from a reader following an investment announcement.
Several kinds of work sit inside a data centre
AWS data-centre careers page illustrates the range at one operator: server operations, engineering operations, infrastructure design, physical security and supply-chain work. It describes electrical, mechanical and HVAC experience as relevant to some roles. This is employer recruitment material, not an independent labour-market study or a vacancy list for the proposed Trinidad facilities.
The page also describes training routes, including a work-based learning programme and an internal programme for existing Amazon staff. Eligibility, location and available places need checking for any advertised opportunity. Their existence does not mean a Trinidad applicant can join from home or that these arrangements apply to EY or Hummingbird.
Use those examples to make your questions more precise. If your background is electrical maintenance, ask how the duties of a facilities role compare with work you already do. If you write software, examine the systems and experience requested for a computing role. There is little value in treating everyone interested in the sector as if they need the same course.
For specialist engineering, keep the limits clear. Reading a job description or taking an introductory AI class does not authorise someone to design a power installation, service live equipment or operate a production network. Any practical work needs suitable training, supervision and permission. A career-planning exercise can begin with public documents and a conversation with a qualified adviser.
Read UWI's programme for what it offers
UWI St Augustine announced on 10 July that it would introduce MASc, MPhil and PhD programmes in AI from September 2026. Its release described technical research, applied problems and responsible AI. The announced application deadline was 17 July, which has passed. The announcement alone does not verify enrolment, completion or placement in a data-centre job.
The MASc programme page covers machine learning, data engineering, intelligent systems and governance, with project work. It lists two years full-time or three years part-time and describes entry requirements and possible qualifying courses. Applicants should confirm the next intake, fees and their own eligibility directly with the university.
That curriculum may be relevant to an AI career, but a prospective student should still compare it with the work they want. Ask which modules and projects address the employer's requirements, what practical experience is available and how it is assessed. Do not assume the word “AI” means every graduate will be prepared for every infrastructure job.
A practical skills check before you spend
The published People First, Problem Second, AI Last approach starts with the person and the problem. Here, the problem is choosing a sensible next step for your working life. You can investigate it with a notebook or spreadsheet before involving an AI tool.
Choose one role that interests you. Save a current public job description from a verifiable employer, including its date, location and application deadline. Record whether it is an actual vacancy, a general description of work or a possible future position. Avoid treating a foreign listing as evidence that the same job exists in Trinidad.
Make three short lists: tasks you can already demonstrate, requirements you could meet with further training, and points you do not understand. Use examples from your own experience, without including confidential employer records. If a requirement says “experience”, ask what evidence the employer accepts rather than assuming a certificate substitutes for it.
Take the unclear points to a course adviser or someone qualified in that occupation. Ask about the equipment and supervision available, how competence is assessed and whether there is a documented placement arrangement. If a provider mentions an employer partnership, check what the partnership actually covers and whether places are guaranteed, competitive or still proposed.
Write down the full commitment before enrolling. Consider tuition, examination or equipment costs, travel, study time and any effect on your current work. Ask for the provider's written terms and take time to understand them. A headline about national investment is not enough evidence to justify a loan or leaving a job.
Set a personal review date that gives you time to get answers. At that point, decide whether you have enough information to proceed, whether a smaller preparatory step would help, or whether you should wait. These are suggested planning steps, not an admissions timetable, a guarantee of employment or a test of someone's suitability for hiring.
If an AI assistant helps compare public job descriptions with course outlines, check every comparison yourself. Require a link to the passage behind each claimed match. Keep identity documents, application records and private workplace material out of an unapproved service. Do not let an automated score decide whether a person deserves a job or training place.
What employers and training providers could publish
A useful skills plan would name the work expected at each stage of a project and explain which assumptions might change. It could show when recruitment is likely, which roles need experienced specialists and where supervised training could help local applicants. Until there is evidence, label these items as plans rather than achievements.
For a training partnership, readers need more than the names of the organisations involved. They could ask how places are funded, who supervises practical work, how participants are selected and what happens if construction or hiring is delayed. Clear information helps an applicant judge the opportunity without having to guess what an announcement means.
Later reporting could distinguish people who enrolled from those who completed training and those who entered relevant work. Publish aggregate results without exposing participants' private information. Explain the period covered and the type of employment counted. These are proposed reporting questions; this article has not verified such results for the projects.
What this review cannot tell you
We have not audited Trinidad and Tobago's workforce, verified the latest financing or construction position, or counted vacancies linked to the proposed facilities. We cannot predict how many local people will be hired or what they will earn. The sources describe proposals, an engineer's concerns, an overseas operator's roles and university programmes, each with a different purpose.
The useful response is to ask for the missing connection between a project and a person's next step. Keep the employer, occupation, required experience and training costs in view. That gives a worker or student something they can check, revise and discuss with an adviser as better information becomes available.
Frequently asked questions
Were 5,000 AI jobs confirmed?
No. The July estimate covered three proposed projects, including steel, and depended on their advancing. It cannot be read as a list of confirmed AI vacancies or local hires. Ask for project-specific recruitment information before using it to make a career decision.
Was Alexander speaking for AWS?
The Express explicitly described his comments as personal and independent of AWS. His technical experience helps explain his perspective, but the interview should not be presented as an employer announcement or a completed national skills audit.
Does an AI degree guarantee a data-centre job?
No. Compare the programme's content and practical assessment with a specific role's requirements. Confirm current admissions information with the university and check employment claims with the named employer. This review does not establish a placement guarantee.
What can I do while project details remain uncertain?
Examine one verifiable role, list the requirements you already meet and ask a qualified adviser about the gaps. Gather written information on costs and practical experience before enrolling. Keep your plan open to revision as project and recruitment information changes.
Sources and update
- Vishanna Phagoo, Trinidad Express interview, 18 August 2026
- AWS technical article and Anton Alexander author biography, 27 October 2025
- Government MoU statement reproduced by TTT News, 11 July 2026
- AWS data-centre careers and training overview
- UWI AI postgraduate-programme announcement, 10 July 2026
- UWI MASc AI programme and entry requirements
Updated 13 October 2026 to clarify the jobs estimate, distinguish personal commentary from workforce evidence, and add role-specific training checks. The original publication date is retained; this update does not report a new project milestone.