Critical thinking
Start with the problem.
Work out the outcome, the constraints and the ways the system can fail. Separate what is true from what is assumed. Question the brief when the brief is wrong.
People who can enter an unfamiliar domain, learn quickly, connect ideas across disciplines and still make something work.
Celeco is early. These are the standards we use while hiring, building, disagreeing and making decisions.
Scroll to enter ↓The problems we take on rarely come with a clean brief or an obvious answer. Every role calls for first-principles thinking, a view of the wider system and the judgement to make good calls.
Critical thinking
Work out the outcome, the constraints and the ways the system can fail. Separate what is true from what is assumed. Question the brief when the brief is wrong.
Systems thinking
Look beyond the feature to the people, incentives, data, hand-offs and exceptions around it. When a problem repeats, turn the answer into a process, tool or piece of shared knowledge.
Judgement & taste
Most decisions arrive before the evidence is complete. Form a view, explain it and let new evidence change it. Taste shows up in work that is clear, useful and cared for.
We want people who want to win, care about being the best and feel the result personally. We should be fiercely competitive about the quality and impact of our work.
Direct that ruthlessness towards the problem, the standard and the pace. Treat teammates, customers and competitors with decency. Help the team win, give credit freely and leave status games behind.
We are a David taking on giants. Speed is the advantage within our control. Our moat is the speed at which we learn, ship and correct.
Make reversible decisions quickly. Build the smallest thing that can produce an honest signal. Within an agreed boundary, act first and explain afterwards. Know when customer trust, privacy, security, safety, legal obligations or significant money make a decision a one-way door.
Long-term value comes from customer satisfaction compounding over time. We earn that one useful decision, one kept promise and one working system at a time.
Stay close after shipping. If people cannot use the system, do not trust it or see no measurable improvement, the work is unfinished.
AI produces passable work in seconds. The details reveal whether a person cared.
Check the awkward edge case. Read the error message. Use the product as the customer would. Notice the broken alignment, the confusing sentence, the missing metric and the assumption nobody tested. Finish the final ten per cent.
Experience is useful. Curiosity lasts longer. Ask good questions, run experiments, learn unfamiliar domains and change your mind when the facts change.
“I don’t know yet, but I’ll find out” is a strong answer when it is followed by action.
Aim to become exceptional at whatever you own. Study the best work you can find. Practise deliberately. Ask sharp people to critique you. Learn the history of the craft as well as its newest tools.
The standard has to be visible in the work, repeated over time.
Early-stage work rarely comes with a complete map. Find the path. Seek context, make sensible assumptions, ask for help before you are stuck and keep moving through ambiguity.
Owners bring others in. They never let the problem disappear into silence because the instructions were incomplete.
We will get things wrong. Say it early, repair the damage and share what you learned. Receive feedback without turning it into a referendum on your ability or intent.
Then change the code, process, checklist, review or behaviour. A repeated mistake usually points to a lesson we failed to turn into a system.
The company is the people spending a significant part of their lives together, chasing a common mission.
We begin by believing each other. Trust survives through honesty, follow-through and early bad news. Admit uncertainty. Keep your word. If circumstances change, say so before someone has to chase you.
Be candid without being cruel. Share context. Give credit. Respect other people’s time. Help without keeping score. Disagree with an idea without diminishing the person behind it.
Each person should stand out in their own right. Say what you believe. Ask the question you think may sound foolish. Admit when someone else has the better answer.
The best work often contains both obsession and fun. Care about our mission and care deeply about something outside work too. Those interests bring perspective and creative range into the room.
If Celeco asks for unusual ambition, speed and ownership, the company owes people an equally serious commitment in return.
We want to build the highest-talent-density company in our category. Every person should raise the room through some combination of raw ability, judgement, curiosity, pace, craft and generosity.
Working together should expose you to better questions, higher standards and ways of thinking you would not have reached alone. A conversation here should be capable of changing your mind or improving your work.
Our preferred company is a smaller team of exceptional people, each with greater scope, stronger support, better pay and real upside.
Money matters. It buys freedom, reduces background anxiety and is one of the clearest ways a company shows what it believes someone’s contribution is worth.
The complete Celeco deal should include inspiring colleagues, difficult and worthwhile problems, real autonomy, rapid learning, excellent tools, strong cash compensation and meaningful ownership.
People here should grow faster, do better work and share materially in what they help build.
Forward-Deployed AI Engineer
Enter a customer environment with an unfinished question and leave behind a working, measurable system. You will move between workflow discovery, engineering, applied AI, evaluation and customer delivery.
Read the role