The right developer for every task.Transparent, scored, auditable.
100-point match score per developer per task. Five factors. Top 5 matches returned with reason bullets. No black box.
The problem
Engineering managers are asked to assign tasks to people in 15-minute sprint-planning slots. The math to do it right - capacity, experience, component familiarity, skill match, current workload - is real. In most teams it gets compressed into 'who's free and who hasn't complained recently.'
- Over-capacity sprints because utilization is a gut check.
- Mismatched tasks - seniors on trivial work, juniors on code they've never seen.
- No audit trail. You can't defend a decision you can't explain.
How it works
Step by step.Nothing behind a curtain.
Capacity availability - up to 25 pts
Full credit if remaining capacity ≥ task SP. Partial (10) if some remains. Zero if over-committed.
Experience-to-complexity match - up to 20 pts
A 4-by-3 matrix of (experienceLevel, taskComplexity). Senior on simple work loses points. Junior on complex work loses points.
Component familiarity - up to 30 pts
30 for 2+ matched components, 20 for 1, 5 for none. The biggest single factor.
Skill match - up to 15 pts
15 if ≥75% of required skills match. 10 if ≥50%.
Workload balance - up to 10 pts
10 below 70% utilization. 5 between 70–90%. Zero above. Total: 100.
What you get
Specific.Not marketing-department specific.
- Top 5 developer matches per task, each with a 100-point score.
- Reason bullets - up to 4 per match, explaining why the score is what it is.
- Risk scoring. Over-capacity flags at >90% utilization (+25 risk points), 70–90% (+10 risk). Risk rolls into the overall recommendation.
- Estimated completion time. Base SP × 0.5 days, multiplied by experience (expert 0.7× → junior 1.4×), by risk (up to 1.5×), and by workload (>80% util × 1.3×).
- Confidence levels tied to data quality and risk, not just score.
Built right
Scored.Sourced.Auditable.
Every match comes with its reason chain - the factors, the weights, the score. Rule-based, not a black-box ML guess. Managers can disagree, override, and see exactly what changed. Implementation depth is best covered live on a call.
Get the architecture deep-dive on a callWorks with
Stop assigning by gut feel.
Private by default. Specific by design. Free for 2 users, forever.