The setting
's gives people exercise and rehabilitation programs for pain in muscles and joints. They use an app and motion-sensing devices from anywhere, and a physical therapist directs the treatment plan.
The active protocol and the agent's retrieved guidance no longer match.
Fast exercise guidance protects member momentum, while protocol fidelity protects clinical judgment.
Who you’d be doing this for
“My renewal deck is due soon, and I need to know the guidance members see matches the care plan.”
Chiamaka Antwi · Benefits Manager · a character in this brief
Sponsors the covered rehabilitation benefit and must explain member outcomes and program reliability during renewal reviews.
What is at stake
Clinicians are rejecting exercise-progression recommendations because the agent retrieves outdated protocol content. You must weigh a fast retrieval correction against regression risk across shared deployments.
Why it isn’t already fixed
Every obvious fix costs something else. That’s the part you’d have to decide.
- clinical fidelity vs response latency
- shared architecture vs client-specific protocol updates
- fast containment vs durable retrieval evidence
- autonomous guidance vs clinician review
Why Sword Health
Camera-guided rehabilitation depends on exercise guidance that stays aligned with clinician-controlled treatment plans.
Written with these in mind
Not your kind of problem? 42 more at Sword Health, or browse every organization.
Where the organization context comes fromThe postings this brief was written from, and the reporting read before it.
The organization context comes from public information. The situation and characters are written for the brief.
Job postings it was written from
Public descriptions of the work at Sword Health. They are listed as sources, not as openings. The lines under a posting are its own words, where it meets this brief.
- Senior Backend Engineer, AI Solutions (Greek-speaking) (opens in a new tab)
- Senior AI Software Engineer (opens in a new tab)
- Senior QA Analyst, AI Solutions (opens in a new tab)
- Senior ML Engineer · no longer posted
- Build agentic LLM systems: design multi-step workflows with tool use, retrieval, and orchestration, and make them reliable enough for clinical settings
- Improve model quality with whatever fits the problem: prompting, retrieval, distillation, or fine-tuning, chosen on evidence rather than habit
- Senior AI Engineer (opens in a new tab)
This is the setup. The work is inside.
Running it puts you in the room: the full situation and its constraints, stakeholders who push back in their own words, and the decisions that are yours to make. What you produce becomes a Day One Plan — work you can show someone instead of describing.
Where similar work appears
Other live briefs connected to this one, and what connects them.
Similar problems at other organizations
Like this one, each of these is about people starting something and never quite finishing it.
Calibration completion fell from 84% to 71% after members resume the app.
You have to weigh a focused client-state correction against a broader change to the shared GraphQL resume contract.
P95 first response has risen from 1.1 to 2.4 seconds after redundant retrieval calls appeared.
You must weigh a faster cache path against session-safe context and answer quality.
Successful grounded CRM tool calls fell to 91.8% after a response-contract change.
You have to weigh a narrow adapter repair against a broader Java compatibility path that consumes the same release window.
The same setting, from another role
Each is set around Sword Health’s Virtual musculoskeletal (MSK) care, written for a different role. They’re separate briefs, not the same scenario.
Five of seven incoming benefits owners couldn't explain where member access questions belong.
You must weigh a focused handover against a broader service that prepares them to defend the benefit at annual planning.
CSMs find unfinished launch steps even when health-plan accounts appear on track.
You decide whether AI can flag a stalled commitment from correspondence or must wait for a CSM to confirm which date applies.
Employers are asking to hold back part of MSK fees when enrollment or outcomes fall short.
You have to weigh a compelling commercial commitment against measures that clinical, legal, and buyers can defend.
Other problems with a similar purpose
In each of these, as in this one, the aim is to make a machine-driven decision dependable and fair at the exact real-world moment a customer is counting on it.
Agents are retrieving older interaction summaries after a customer updates a preference.
You must weigh a ranking correction against retrieval latency and safe rollout coverage.
Fourteen percent of Qwen requests now exceed three seconds after the container build.
You must weigh faster batching against the context available to each answer.
Duplicate medication events appear in 6.8% of amended-record cases and inflate treatment timelines.
You have to weigh fast suppression against losing real repeat prescriptions.