How engineering and HR leaders design global teams without breaking delivery.
The companion report to the Hacking Borders podcast — a research dossier for TA, People, and engineering leaders at Series A–C tech companies scaling distributed engineering through the AI era. Built from structured operator conversations with TA, People, and engineering leaders at US growth-stage and mid-market technology companies.
The bottleneck moved. The org charts didn't.
Mid-market technology organizations spent the last decade optimizing for hiring velocity. The data from 2025 is unambiguous: the constraint has shifted. The organizations compounding execution are not the ones with the most engineers. They are the ones whose management systems, onboarding pipelines, and AI-integrated workflows scale at the same speed as headcount.
The DORA conclusion is the operational thesis of this report. AI is not a productivity tool that lifts every organization equally. It is a diagnostic that exposes which organizations have built operating systems capable of absorbing acceleration — and which have not. The companies winning the next decade are converting that diagnostic into structural change, before the gap becomes commercially decisive.
Four shifts this report documents
What fails first — quietly — before organizations realize they have stopped scaling.
Scaling failures rarely announce themselves. By the time leadership sees the symptoms — missed quarters, attrition spikes, delivery slippage — the structural causes have been compounding for two to four quarters. The sequence is predictable.
The cascade is causal — each stage produces the next. Manager saturation degrades onboarding; degraded onboarding inflates coordination cost; coordination cost surfaces as decision fog; decision fog slips roadmaps; slipped roadmaps drive senior attrition. The reorg at the end treats whichever stage is most visible to the executive team — usually stage 5 or 6 — rather than the root cause at stage 1. The cycle restarts. Most organizations have seen this loop more than once and not recognized it as a loop.
AI is not the productivity revolution. It is the diagnostic.
The dominant 2024–25 narrative held that AI would compress engineering org sizes and accelerate output across the board. The 2025 DORA findings — the largest controlled study of AI-assisted engineering to date — show a more specific reality: AI amplifies the system it operates inside. The implication is structural, not technological.
The McKinsey 21% workflow-redesign finding is the most underweighted number in the AI literature this year. AI is helping individual developers ship more code. It is not yet helping teams ship more product — because most organizations are layering AI on top of legacy workflows rather than redesigning around it. The bottleneck moved upstream: code production is no longer the constraint. Code review, integration testing, decision flow, and cross-team handoff are now the constraint — and those are organizational systems, not engineering tools.
Three organizational archetypes after AI
How modern engineering organizations structure for scale.
The operating model that carries an engineering org from 50 to 200 engineers — functional teams, weekly all-hands, Slack as decision substrate, managers as coordination glue — breaks at 300+. Organizations either redesign before the breaking point or pay 3–6 quarters of remediation after it.
| Archetype | Coordination mechanism | Effective range | How it breaks |
|---|---|---|---|
| A · Heroic | Individual memory; Slack-as-decision-store | Up to 80 engineers | Founder/EM saturation. Onboarding depends on availability of one person. |
| B · Functional | Manager hierarchy; weekly sync cadence | 80–250 engineers | Coordination tax surge. Meetings expand to absorb decision throughput. |
| C · Pod-based | Outcome-aligned pods; documented decision rights | 150–800 engineers | Platform under-investment. Pods reinvent infrastructure in parallel. |
| D · Network | Handbook-first async; codified decision rights | 500+ engineers, multi-time-zone | Documentation debt. Operating model only as good as its substrate. |
Sijbrandij's framing maps directly to the operating model progression. Archetypes A and B can function with synchronous coordination because the org is small enough or co-located enough that meetings absorb the decision load. Archetypes C and D cannot. Pods and networks require an async substrate — a single source of truth that allows decisions to flow without synchronous gating. GitLab's 2,700+ page public handbook is the operationalized version of that principle.
Distributed engineering is no longer a cost strategy. It is the operating model.
The framing inherited from the 2010s — offshore engineering as a cost lever — has been overtaken by the operating reality of 2025. The frontier is coordination integrity, time-zone alignment, and operating model compatibility across borders. Cost is now downstream of the model decision.
Atlassian's principle captures what the Stack Overflow data implies. When a third of engineers work remotely, physical location is no longer the design variable. Time-zone overlap is. The Atlassian State of Teams 2025 report quantified one downstream cost of getting this wrong: teams now spend over 25% of their workweek searching for information. That is not a tooling failure. It is the operating signature of a distributed organization without an async substrate to support it.
The two distributed engineering models
- 8–12 hour time-zone gap between HQ and engineering staff
- Coordination via overnight handoff documents and morning calls
- Engineers operate outside HQ pods — "the offshore team"
- High substitutability assumption · low retention
- Decisions escalate to HQ before being made
- Optimized for cost-per-engineer
- Breaks down at organizational stage 3+ when coordination cost exceeds savings
- 0–2 hour overlap with HQ (Mexico ↔ US Central/Mountain)
- Real-time pair work, review, and decision flow
- Engineers operate inside HQ pods · not a separate unit
- Long-tenure retention · full integration
- Decisions made at the pod level
- Optimized for shipped-value-per-engineer
- Compatible with Archetype C and D operating models
Time-zone overlap with US HQ — by engineering region
Why Mexico is the structural answer for North American mid-market
Five organizational maturity stages — and where you sit determines what fails next.
The Hacking Borders maturity model classifies engineering organizations along five stages that progress not by headcount but by operational system depth. Each stage has a predictable breaking point and a specific set of investments that unlock the next stage.
Breaking point: senior leaders saturate.
Breaking point: coordination tax outruns capacity.
Breaking point: documentation drift; systems decay without ownership.
Breaking point: global expansion exposes async gaps.
Breaking point: leadership transitions.
The six-dimension scorecard
Stages are useful for orientation. Operational change requires diagnosis at the dimension level. The Hacking Borders Assessment scores organizations across six weighted dimensions — each measurable, each improvable, each tied to a specific failure mode.
| Dimension | What it measures | Failure mode if low |
|---|---|---|
| Management Scalability | Span discipline, 1:1 cadence integrity, manager coaching capacity, skip-level depth. | Senior attrition, coaching collapse, decision-by-availability. |
| Organizational Alignment | Codified decision rights, written-first culture, RFC discipline, cross-team contracts. | Roadmap drift, recurring "who owns this?" friction. |
| Capacity Sustainability | Onboarding maturity, time-to-first-commit, knowledge documentation, mentorship density. | Hiring-as-tax, senior bandwidth drain, year-one retention failure. |
| Distributed Execution Readiness | Async maturity, time-zone integration discipline, written communication norms. | Distributed teams operate as second-class citizens; boundary failures. |
| AI Adaptability | Workflow redesign depth, review capacity scaling, AI-augmented onboarding, governance. | AI raises throughput without raising shipped value; instability climbs. |
| Operational Infrastructure Maturity | Platform investment, developer experience, paved roads, system of record discipline. | Pods reinvent infrastructure; platform debt compounds. |
Benchmark · Where mid-market engineering organizations actually sit
The distribution below is modeled from public engineering performance datasets — DORA 2025, McKinsey State of AI 2025, LeadDev Engineering Leadership 2025, and the Atlassian State of Teams 2025 — mapped onto the Hacking Borders five-stage framework. The Hacking Borders Assessment will refine these benchmarks as Season 1 survey responses accumulate. Use this as your starting reference point; the assessment locates your specific organization within the distribution.
Benchmark · Six dimensions, six benchmarks
The six assessment dimensions are not abstract. Each maps to a publicly measured benchmark you can compare your organization against today. The reference numbers below are what the survey would reveal — drawn from the same external research that calibrates the Hacking Borders model.
| Dimension | Benchmark reference | What this tells you about your organization |
|---|---|---|
| Management Scalability | 12.1 average direct reports per manager today; 46% of engineering leaders at moderate-to-critical burnout. | If your engineering spans exceed 10, you are in the burnout zone Gallup measured directly. Coaching collapses past that point. |
| Organizational Alignment | 25% of the workweek is now lost to searching for information across tools and teams. | The cost of unclear decision rights, scattered docs, and tribal knowledge. Top quartile orgs have eliminated most of this through written-first culture. |
| Capacity Sustainability | Median onboarding ramp: 2–3 weeks to first commit; 3–6 months to full productivity. Top quartile: 1–3 days TTFC. | If you don't track time-to-first-commit, your onboarding is silently degrading. Each hire pays the tax. |
| Distributed Execution Readiness | 32.4% of engineers globally now work fully remote (45% US). Hybrid is dominant — but only 15% of orgs operate Archetype C/D async substrate. | The gap between "we have remote engineers" and "we operate distributed engineering" — most orgs do the first, very few the second. |
| AI Adaptability | 90% of devs use AI assistance (DORA). 84% use or plan to use AI (Stack Overflow). Only 21% of organizations have redesigned workflows around AI (McKinsey). | If your workflows are unchanged after AI adoption, you are in the 79% cohort. The 2.8× redesign delta (55% vs 20%) separates the top 15% from the rest. |
| Operational Infrastructure Maturity | DORA 2025: 16.2% of orgs achieve on-demand deployment; 43.5% have lead time over one week. | The gap between elite and median is platform investment. Below 10% of engineering headcount on platform, pods reinvent infrastructure. |
Six operational moves that compound — most can be started this quarter.
None of the six require new headcount, new capital, or a different talent market. All of them require leadership willing to invest in operating systems rather than hire around the symptom.
The operator network behind Hacking Borders.
The podcast
Hacking Borders is a podcast where senior HR and engineering leaders share the real decisions, failures, and systems behind scaling global engineering teams. No theory. No fluff. Each episode is a 30-minute conversation between Tony Le and one guest — operator-level stories from people who have actually done it.
Season 1 runs June–November 2026, sponsored by CodersLink. The format is video-first, released across Spotify, Apple Podcasts, YouTube, and LinkedIn.
The host
Tony Le is a global talent executive and 3× Talent100 winner. He has led recruiting at Mission Cloud, Contentful, Getaround, and IAC, and serves on the HIGHER Community Advisory Board. He has built and scaled engineering recruiting functions through hypergrowth across multiple Series A–C technology companies.
The Hacking Borders research program — including this report and the companion Engineering Capacity Scalability Assessment — is informed by structured conversations Tony hosts with each guest.
The operator panel
The Hacking Borders research program draws on structured conversations with senior TA, People, and engineering leaders at mid-market and growth-stage US technology companies. Each Season 1 episode features one operator-level guest in conversation with Tony Le on the real decisions, failures, and systems behind scaling global engineering teams. Guest names and affiliations are disclosed at episode release; this report synthesizes the cross-conversation patterns rather than attributing claims to individuals.
References
| Source | Focus area |
|---|---|
| DORA — State of AI-Assisted Software Development 2025 | AI amplification thesis; team archetype data; deployment frequency benchmarks |
| McKinsey — State of AI 2025 | Workflow redesign correlation; 2.8× high-performer redesign delta (55% vs 20%); only 21% of organizations have redesigned workflows |
| DORA — State of AI-Assisted Software Development 2025 | 90% AI adoption among developers; 80%+ report productivity gains; AI as system amplifier thesis |
| Stack Overflow — Developer Survey 2025 | 84% AI adoption (51% daily); trust decline 46% (up from 31%); 32.4% fully remote, 45% in US |
| Atlassian — State of Teams 2025 | 25% of workweek lost to information search; "time zones, not zip codes" principle |
| Gallup — Q3 2025 Span of Control | 12.1 average direct reports (up from 10.9 in 2024); ~50% increase since 2013 |
| SHRM — 2025 Talent Trends | 69% of orgs report difficulty filling roles; top barriers: low applicants (51%), competition (50%), candidate ghosting (41%) |
| LeadDev — Engineering Leadership Report 2025 | 46% combined moderate-to-critical burnout (22% critical + 24% moderate); 65% report expanded scope |
| GitLab — Public Handbook | Handbook-first operating model at 2,000+ people across 70+ countries |
| IDB — Nearshoring opportunity for LATAM | $78B annual export gain; Mexico captures $35.3B (45% of regional opportunity) |
| Gartner — Worldwide IT Spending Forecast 2026 | $6.15T total IT spending 2026; IT services segment $1.87T; nearshore as fastest-growing sourcing segment |
| CodersLink — Mexico Tech Salaries Report 2026 | n=10,246 verified responses, 36 roles, 32 states; cost and compensation benchmarks |
| McKinsey — Redesigning the Technology Workforce for the Agentic AI Era (2025) | GitLab all-remote handbook-first model as case study for agentic-AI-era workforce design |
| Gartner — IT Outsourcing & Nearshore Forecast 2025 | Codified decision-rights model correlation with distributed engineering performance |
| INEGI — ENOE 2025 | Mexico occupational data; ~700,000 software developers |
| BLS — OEWS 2024 | US senior engineer compensation benchmarks for nearshore cost-savings comparison |
| Hacking Borders — Season 1 Operator Panel | Structured conversations with TA, People, and engineering leaders at US growth-stage and mid-market technology companies |
A decade of building dedicated Mexico engineering teams.
CodersLink has spent 11 years operationalizing the distributed engineering model documented in this report. 150+ companies across Software, FinTech, HealthTech, Logistics, and Enterprise SaaS have scaled their engineering organizations through the CodersLink network and delivery model.
What clients say
