Most neurotypical people have a relatively flat cognitive profile: if they score average on spatial reasoning, they tend to score average on verbal reasoning and working memory too. Their abilities cluster around the mean across most dimensions.
Neurodivergent people frequently have a spiky profile — extreme peaks in some cognitive areas and significant troughs in others. This is not a metaphor. Cognitive assessments of autistic individuals, for example, routinely show performance at the 99th percentile on tasks like pattern detection or systemizing, combined with performance at the 10th percentile on tasks like processing ambiguous social language or managing unstructured time.
What the Peaks Look Like in Practice
- An engineer with ASD who can hold the entire dependency graph of a microservices architecture in working memory and immediately identify which service will fail under a specific load pattern.
- An engineer with ADHD who, during a hyperfocus episode, rewrites a critical performance bottleneck in a single overnight session — producing work that would take a neurotypical engineer a week.
- An engineer with Dyslexia who thinks in three-dimensional spatial models and designs infrastructure topologies that are architecturally elegant precisely because they visualize the system as a physical structure rather than a text description.
What the Troughs Look Like in Practice
- Forgetting to update a ticket in Jira because the context-switching required to move from code to admin work is cognitively expensive.
- Struggling to produce a coherent verbal summary in a standup meeting, not because they don't understand the work, but because converting deep technical knowledge into spontaneous spoken language is genuinely hard.
- Missing a deadline because the task was communicated ambiguously and they interpreted it differently than intended.
The Management Mistake
The instinctive managerial response is to focus on the troughs. "Alex needs to get better at writing status updates." "Jordan needs to improve their communication in meetings." This approach attempts to flatten the spiky profile — to make the neurodivergent engineer perform like a neurotypical one. It almost never works, and it burns out the engineer in the process.
The inclusive approach is the opposite: capitalize on the peaks and engineer around the troughs.
- Don't force the engineer to be good at administrative overhead. Give them tooling, templates, or a buddy system that handles it.
- Don't force them to perform verbal fluency in meetings. Let them contribute via written async channels where their communication is often exceptional.
- Do give them the complex, ambiguous technical problems that nobody else wants to touch. That is where their architecture pays off.
The analogy in systems design: you don't run your GPU on general-purpose CPU tasks. You route the right workload to the right processor. Managing neurodivergent engineers is the same principle applied to human cognition.
Interviewers sometimes ask: "How do you manage underperformance on your team?" A strong answer acknowledges that apparent underperformance is often a mismatch between the person's cognitive profile and the environment or task assignment — not a capability deficit. Describing how you identified a spiky profile, adjusted task allocation, and saw performance improve demonstrates both management maturity and systems thinking.