01 The Challenge
Spectocor — a medical-device company building a heart-analytics device data platform —
had spent $15 million building out a distributed development program and was
producing no results. The platform that was supposed to ingest, process, and surface device
telemetry to clinicians was stuck. The codebase had grown, the headcount had grown, the spend
had grown, but functionality had not.
The structure: a 160-developer offshore program spread across five vendor
groups — two in China, three in India — coordinating across time zones, tech-stack opinions,
and divergent codebases on top of an unclear architectural vision. The leadership team needed
an outside diagnosis: what is actually wrong, and what does “fixed” look like?
02 The Diagnosis
Leopard Data engaged as Solutions Architect on top of the 160-developer program.
The first deliverable was not code — it was a 50-page diagnostic document
tracing why the investment was failing to produce results. The document covered architectural
debt, communication overhead, scope of accountability, code-quality patterns across the five
vendor groups, and the fragmentation that comes from having no single technical owner with
end-to-end visibility.
On top of that diagnosis we wrote a short executive brief — a leadership-readable
version that gave the company something concrete to work toward. The brief named the
root causes, presented options with cost and risk profiles, and proposed a path forward
grounded in the business outcome instead of the headcount.
03 The Pivot
The recommendation that came out of the diagnostic: scale was the wrong lever. A
160-person distributed program had produced fragmentation, not throughput. The fix
was the inverse: a small, senior, onshore team with end-to-end ownership.
Leopard Data led the pivot. The 160-person program was wound down. We hired
five senior onshore full-stack developers — not as a near-shoring cost
play, but as a deliberate bet that five engineers who could own a system end-to-end would
outperform 160 who couldn’t. Allan personally ran every interview, made every hire, and
onboarded the team before the engagement closed.
04 The Outcome
- Project recovered. The platform that had stalled under a 160-developer program began shipping under the new team.
- Diagnostic document and executive brief became the ongoing decision framework for the platform’s technical direction.
- Five senior full-stack developers personally interviewed and hired by Leopard Data before handoff — the team that owned the system going forward.
- Demonstrated, in production, that small-senior-onshore can outperform large-distributed-offshore on the right kind of problem.
05 Tech Stack
.NET
Hadoop
SQL Server
Azure (Hybrid Deployment)
Big Data Pipelines
Heart-Analytics Device Telemetry
