Addaly is in open beta. Things will change, and AI answers can be wrong — check anything that matters.

How people actually get in

Working In AI · lesson 5 of 7 · 8 min

The uncomfortable base rate

Applying cold to public postings is the most common strategy and the weakest one. A single AI-adjacent posting at a known company can draw several hundred to a few thousand applicants, most auto-screened. Sending two hundred applications and hearing back twice is not bad luck; it is the ordinary output of that method.

The methods that work all have the same shape: a human sees evidence of your work before, or instead of, an application landing in a pile.

The four routes that actually carry people

Sideways from an adjacent job. The most reliable and least discussed. Get any job at an organisation that has data — support, operations, QA, finance, field work — and become the person who answers questions with numbers. You already have context nobody outside can buy, you are already trusted, and internal moves skip the entire screening machine. Large numbers of working analysts and data engineers arrived this way.

Referral. A referred application is read by a person, usually within a day. This does not require a network of executives. It requires being known by ten or fifteen people who do this work, which is a matter of months of showing up in the same places and being useful.

Small companies. A company of thirty reads its applications. A company of thirty thousand does not. Your careful cover letter is only ever read at the small one.

Paid work first. A short contract, a freelance job, a three-month project. Once someone has paid you, you are experienced, and every subsequent conversation starts differently.

The market you are in changes the shape

India. Campus placement and the large service companies remain the widest entry route, with aptitude and coding tests up front. Getting in at scale then moving internally into a data team is a completely legitimate path and a common one. Notice periods of 60 to 90 days are normal and affect how employers plan hiring.

Nigeria and Kenya. Community is the hiring channel. Developer groups, talent programmes, public work shared where people can see it. A large share of good jobs are remote contracts with foreign companies, which means your competition is global and your evidence must be public.

Latin America. Nearshore staffing firms placing engineers with US companies are a major employer. The real gate is often spoken English in meetings rather than technical depth.

Europe. Structured, slower processes, several rounds, formal references. Visa sponsorship narrows the list of employers sharply, and in Germany, France and Spain a working level of the local language quietly decides many mid-level roles.

United States. Referral-heavy, take-home exercises common, brutal application-to-interview ratios at known names. Non-obvious employers — hospitals, insurers, logistics firms, universities, state agencies — hire steadily with far less competition.

Southeast Asia. Regional scale-ups hire in cohorts, often with structured graduate intakes that behave more like the Indian model than the American one.

What to do this month

Pick six companies you could plausibly join, not sixty. Find one person at each who does the work. Read something they wrote, then send four sentences: what you built, one specific question about their problem, no request for a job. Some will not reply. Two or three will.

That is a worse-feeling and much higher-yield use of an evening than another fifty applications.

Before you move on

Meera has sent 200 applications through job boards over six months and received two replies. She writes clearly and her projects are solid. Which change is most likely to help?

Pick the one you would defend. Nobody sees your answer.

No ads. No data sale. No public scores on people. Ever.

© 2026 Addaly