Operating model & engagement

Capacity and skills coverage

Resolve the specific skill gap and see the schedule and cost consequence.

Current decision

3 skill gaps still need a decision

Delivery 30 weeks (10 beyond plan); monthly cost $480k, including $0k in added capacity.

Start with ML Engineering; choose hire, contract or upskill for each gap.

Same instrument · 3 industries choosing a scenario replaces its inputs; save or share your current changes first

Team

30 FTE

7 FTE short in skills

Delivery

~30 wk

+10 wk vs plan

Monthly cost

$480k

base team

Unresolved gaps

3/3

Over allocated skills

Skill utilization · demand ÷ capacity

ML Engineering

demand 9 · capacity 6 · 150%

Data Engineering

demand 7 · capacity 5 · 140%

MLOps / Platform

demand 4 · capacity 4 · 100%

balanced

Delivery / PM

demand 4 · capacity 5 · 80%

+1 slack

Domain SME

demand 5 · capacity 6 · 83%

+1 slack

QA / Eval

demand 6 · capacity 4 · 150%

Every bar uses the same scale. The tick marks 100% utilization; bars beyond it show overload, with the exact percentage below. Hire = +6 wk / $18k·FTE · Contract = +1 wk / $28k · Upskill = +4 wk / $8k (draws on slack).

Bottleneck: ML Engineering

Capacity decisions remain for ML Engineering, Data Engineering, QA / Eval. Contract is fastest to the date, upskill is cheapest but leans on the slack in Delivery and SME, hiring is permanent but adds six weeks. Pick per constraint, not per habit.

If you act on this · the call → expected lift → how you'd measure it

The call

Resolve capacity gaps by skill, not by generic headcount.

Expected lift · illustrative

Improves delivery confidence by matching the work to the actual capabilities required.

How you'd measure it

Skill utilization, open gaps, delivery date impact, monthly cost, time to productive capacity.

Steering committee takeaway: Thirty people do not equal thirty usable delivery units. Capacity fails by skill, not by headcount.

Resume echo, a direct mirror of the 31-resource AMEX intelligence mapping; the most personal instrument on the site.

How this is built

Utilization = demand ÷ effective capacity per skill. Resolving a gap adds its shortfall as capacity; delivery slip = the worst gap's overflow (unresolved) or its resolution lead time (resolved). Monthly cost = base team + Σ(gap × resolution rate).

Stack: Next.js (static) + shared design system; deterministic client side.

Limitations: this is a deterministic planner. Real capacity planning would require availability data, role definitions, location constraints, ramp time, vendor constraints, and delivery priorities.

Method, evidence and limitations

Prefer to read? The two minute case study · problem → approach → metric → outcome

Problem

AI portfolios require the right mix of ML engineering, data engineering, platform, delivery, domain, and evaluation capacity. A raw headcount view hides the bottlenecks that delay delivery.

Approach

The planner maps demand and capacity by skill pool, highlights overallocation, and models the effect of hire, contract, and upskill actions on cost and schedule.

Why this way

This connects AI delivery plans to workforce strategy, capacity planning, budget, schedule confidence, and execution risk.

The metric

Utilization versus target; skill coverage gaps against demand.

The trade-off

Hiring is slow and costly; contracting is fast but thin; upskilling is sticky but takes time.

Outcome

A hire/contract/upskill call per gap, with the date and cost impact.