We are building the team that builds the future of enterprise AI.
A practice you can actually own.
The work in front of this firm is the most consequential work in the technology industry right now. Real enterprises — banks, hospital networks, pharmaceutical companies, carriers — are putting agent systems into workflows that matter. The question is who builds them well.
We are an AI-native practice with a sixteen-year operating heritage behind it. That combination is unusual. It means we ship — and we get to work on programs that consultancies cannot reach and startups cannot survive. It also means we operate with the discipline of a firm that has been delivering enterprise programs for a decade.
The team is intentionally small. Senior throughout. Every hire is load-bearing. If you join, you will work on a single program, with a senior team, under a partner who is on the floor — not in the corner office.
Seven roles. All senior. All practice-leading.
-
Architect and ship multi-agent systems for tier-1 enterprise clients. Evaluation-first development. End-to-end ownership: design, build, operate.
Apply for this roleProcess- 01 Intro conversation with a Partner — 45 minutes.
- 02 System-design discussion grounded in a real engagement we have shipped.
- 03 Two technical sessions with the Build practice.
- 04 Final round: time with the Office of the President.
-
Own the evaluation harness as a first-class artefact across engagements. Codify the methodology that makes our work auditable, repeatable, and worth defending in front of a regulator.
Apply for this roleProcess- 01 Intro with the Build practice lead.
- 02 Methodology discussion using a past engagement as the case.
- 03 Two working sessions with Engagement Directors.
- 04 Final round: Office of the President.
-
Run the bridge between frontier research and production agent systems. Translate evals from research papers into the harness our practice uses. Stay close to model labs without forgetting we ship.
Apply for this roleProcess- 01 Intro with the Build practice lead.
- 02 Research-deep-dive on a paper you select.
- 03 Engineering session with a senior on the Build team.
- 04 Final round: Office of the President.
-
Lead AI-portfolio engagements with CIOs and Chief Transformation Officers. ROI modelling, sequencing, capability diagnostics. The strategy register of our practice.
Apply for this roleProcess- 01 Intro with the Define practice lead.
- 02 Engagement walk-through using one of your past programs.
- 03 Two client-style sessions with the Define partners.
- 04 Final round: Office of the President.
-
Run a single major engagement end-to-end. Sponsor relationships, program governance, the rhythm of demos and steering. You are the person the executive sponsor calls first.
Apply for this roleProcess- 01 Intro with an Engagement Director and a Partner.
- 02 Program walk-through using one of your past engagements.
- 03 Two sessions with Define + Build leads.
- 04 Final round: Office of the President.
-
Build out the Operate register. Telemetry, cost & capacity, the operational wrapper around shipped systems. The discipline of an SRE practice paired with the empathy of a customer-success function.
Apply for this roleProcess- 01 Intro with the Operate lead and a Partner.
- 02 Operations-deep-dive on a past system you ran.
- 03 Two sessions with Build partners.
- 04 Final round: Office of the President.
-
Own the data-and-retrieval layer underneath agent systems. Chunking, embedding strategy, context-management, evaluation gates over retrieval. The plumbing that determines whether agents work in production.
Apply for this roleProcess- 01 Intro with the Build practice lead.
- 02 Engineering deep-dive on a retrieval system you have shipped.
- 03 Two technical sessions with the Build team.
- 04 Final round: Office of the President.
Four rules, plainly stated.
- 01
Senior-only, across the board.
We hire senior practitioners. No analyst tier. The lowest seniority on the floor is mid-level engineer with five years of production experience. We do not run a graduate program.
- 02
Practical, not algorithmic.
Our loop is grounded in actual past engagements. We do not run live coding puzzles. We do not optimise for LeetCode. We optimise for whether you can ship and explain a system to an executive sponsor.
- 03
Two rounds with partners.
Every hire meets two partners in their loop. The senior team makes the call. Decision within seven business days of the final round, on the record.
- 04
Compensation, plainly stated.
Bands are published on the role description. We do not negotiate against expectations we have not disclosed. The number is the number, and it reflects the seniority and the geography of the work.
Don't see your role? Reach out anyway.
If you have built and shipped an agent system that matters, or you have led an enterprise AI program from strategy through operation, we will make time for a conversation. Write to the Office of the President directly.
Write to info@atsmeridian.ai