About Mosaic Conseil
Mosaic Conseil is a boutique remote consultancy dedicated to rebuilding business artifacts. We operate as a small workshop where we take apart three year plans, pricing grids, and diligence files to put them back together based on the evidence found in the invoice ledger. Our process is personal and focused. We do not produce slide decks or work on standing retainers. Instead, we return the same artifact you sent, redrawn by hand and supported by a written argument. Every file carries two names on the cover: one person who builds and one who argues against what was built. We keep our workshop small to ensure that the people who do the work also answer the phone. We focus on four specific areas: financial services, industrial networks, growing software companies, and healthcare systems. Our partner is a major development in these sectors, and we are expanding our team to include a Junior Prompt Engineer who can help us refine our analytical tools.
What you would do
As a Junior Prompt Engineer, you will support our core analysts by crafting and refining instructions for large language models to assist in the reconstruction of business artifacts. This is not about generating generic content, but about using automated tools to audit complex datasets and financial models. You will work within our builder and challenger framework. One day you might be building a prompt to extract specific anomalies from a healthcare system ledger. The next day, you will act as the challenger, testing a colleague's prompts to see where they fail or produce inconsistent results. This role is fully remote, allowing you to focus on deep, analytical tasks without the interruption of traditional office life. You will be responsible for documenting the logic behind every prompt, ensuring that our automated analysis is as rigorous and evidence based as our manual work.
Helpful background
We are looking for individuals who value intellectual honesty and clear communication. Since this is an entry level position, we do not expect years of professional experience, but we do look for certain qualities. A background in linguistics, philosophy, or basic data analysis can be very useful. You should have a natural curiosity about how instructions are interpreted and a willingness to iterate on a single task until it is precise. Comfort with basic logic and a high degree of literacy in English are essential, as our work relies heavily on written arguments. If you have experimented with large language models through personal projects or university coursework, that experience will serve you well. We value those who can think critically about why a model provides a certain answer and how to adjust the input to reach a more accurate outcome.
How we work
We work in a quiet, remote environment that prioritizes deep focus over constant communication. At Mosaic Conseil, there are no meetings that were not scheduled in advance and no corporate theater. We rely on the builder and challenger model to maintain our standards. This means that for every piece of work you produce, another member of the team will be assigned to find its flaws. This is not a personal critique but a necessary part of our workshop process to ensure every assumption is tested before a file closes. We handle our work on paper and back it with a telephone number that answers for ninety days after a project is finished. This level of accountability requires us to be disciplined and focused on the evidence at the floor of the business.
What the role offers
This is a permanent, full time position that is entirely remote. We offer a work environment where your autonomy is respected and your primary responsibility is the quality of your output. Because we avoid unnecessary meetings and standing retainers, you will have the time to engage in the kind of deep work that is often missing in larger firms. The position is paid, and the compensation is competitive for the role and the current market. You will gain experience in high level business analysis across financial services and industrial networks, working directly on artifacts that influence major decisions. We provide the opportunity to learn a rigorous method of auditing and reconstruction that values evidence above all else.
Our workshop remains small on purpose. By staying independent and focused on the evidence, we ensure that every member of the team contributes to the final artifact. We look for thinkers who are ready to move away from superficial reports and toward the discipline of precise, written arguments.