Easy warm-up
Start with your experience and a small technical exercise so you can settle in and think out loud.
Founder-led · live 1:1 · 90 minutes
You don’t need another course. You need an engineer in your corner.
You already work with modern software, cloud, and AI. This relaxed, mentor-led interview gives us a clear baseline—without the pressure or performance theater of a hiring loop.
One paid session. No long-term commitment. Your rate is confirmed before scheduling.
MITElectrical Engineering & Computer Science
Nearly 20Years in engineering and technical leadership
1,000+Hours of live 1:1 instruction
5.0 · 250+Rating and reviews
The goal is not to judge whether you know enough technologies. It is to discover how you think when the answer is not already in front of you—then turn that evidence into the right training sequence.
I will watch how you frame ambiguity, read code, debug, implement, test, reason about systems, communicate tradeoffs, and evaluate AI suggestions.
You will leave knowing what is already strong, what is holding you back, and what to work on first.
Edward Mabonga
Founder, Pristone Academy
The 90-minute experience
Four rounds. One honest picture.
You will work aloud through four interview-style rounds with Edward. The pace adapts to you: support is available when you need it, and deeper backup questions are ready when you move quickly.
Fibonacci coding
Read, correct, implement, and test a familiar problem. Backup questions raise the ceiling if it feels easy.
System design
Build an architecture from requirements, then respond to changing constraints and bonus scenarios.
AI judgment
Use and challenge AI advice, verify assumptions, and show that you remain the decision-maker.
Relaxed by design: this is interview practice with a mentor in your corner, not a pass-or-fail gate. Rounds 0–2 establish your independent baseline; AI is introduced deliberately in Round 3 to assess judgment, not prompt fluency.
The interview loop
Four rounds. One clear picture.
The questions are deliberately approachable on the surface. The signal comes from how you clarify, reason, test, adapt, and explain—not from recognizing a trick.
Round 0 · Easy warm-up
Start from familiar ground
Settle into the session before the difficulty rises.
What happens
You will explain a project or technical decision you know well, then work through a short, low-pressure code exercise.
The purpose
- Get comfortable thinking aloud with Edward.
- Establish how you clarify a small requirement.
- Create an early success before the adaptive rounds begin.
This is intentionally easy. It is a calibration round, not a trap.
Round 1 · Coding and reasoning
The Fibonacci diagnostic
Read first. Explain second. Change the code only after you have a model.
Starting code
function fibonacci(n) {
if (n <= 1) return 1;
return fibonacci(n - 1)
+ fibonacci(n - 2);
}Your task
- Walk through the behavior for inputs 0 through 5 without editing.
- State the contract you think the function should have.
- Implement a correct version and explain your choices.
- Write the tests that would make you confident in it.
- Discuss what changes when inputs become large or requests become frequent.
If the baseline is easy, Edward will introduce deeper follow-ups on scale, exactness, performance, and production use. You will not be held back by the starter question.
Round 2 · System design
Document answers you can trust
Architecture begins with questions and constraints—not boxes and vendor names.
Design prompt
Design a service where users upload documents, ask questions about them, and receive answers with citations to the source material.
The Pristone interview loop
1 · Clarify2 · Constrain3 · Estimate4 · Contract
5 · Design6 · Attack7 · Operate8 · Defend
Be ready to explain
- The requirements, non-goals, and rules that must never be violated.
- One scale estimate and the decision it changes.
- The API, data model, major components, and numbered data flow.
- How failures are detected, contained, recovered, and prevented from recurring.
- Which tradeoff you made and what evidence would change your decision.
Bonus constraints are introduced live if you fly through the baseline. Naming a RAG stack is not enough; the diagnostic tests the engineering reasoning underneath it.
Round 3 · AI engineering judgment
Use AI without surrendering ownership
The skill is not getting an answer. It is knowing whether the answer deserves trust.
Your task
Evaluate AI-generated advice about the system you designed. Decide what to accept, reject, test, or rewrite—and explain why.
Be ready to discuss
- How you verify technical claims and hidden assumptions.
- Security, privacy, hallucination, evaluation, and cost.
- Where a human decision must remain in the loop.
The round measures engineering judgment with AI, not prompt tricks or memorized terminology.
What you keep
The session produces a plan—not just a score.
Even if you do not continue into mentorship, the diagnostic is designed to leave you with a useful account of where you are and what should happen next.
Capability map
Observed strengths and gaps across decomposition, coding, debugging, testing, systems, communication, AI judgment, and independence.
Evidence, not labels
Specific behaviors from the session, including where you were independent, where prompting helped, and which risks were otherwise easy to miss.
Prioritized sequence
The smallest set of development priorities most likely to create meaningful progress—ordered rather than dumped into a generic curriculum.
First-phase roadmap
A recommended six-to-eight-week opening phase, including practice, independent work, and evidence that the new capability transfers.
Cadence recommendation
An honest recommendation for session frequency and scope based on your goal, baseline, and rate of progression.
A go / no-go decision
A clear view of whether long-term mentorship is the right investment—or whether a narrower sprint or independent plan makes more sense.
If we continue
A mentorship that gets harder as you get stronger.
The exact sequence is earned from the diagnostic. The common arc moves from strong foundations toward independent performance under ambiguity and interview pressure.
Engineering foundations under real work
Problem decomposition, code reading, debugging, implementation, testing, data structures, and deliberate practice.
Reliability and system behavior
APIs, data, concurrency, failure boundaries, observability, and the difference between toy correctness and production confidence.
Architecture and technical judgment
Requirements, constraints, design alternatives, tradeoffs, distributed systems, cloud, and production AI workloads.
Independent and interview-ready performance
Unfamiliar problems, mock interviews, design communication, project explanations, and progressively less guidance from Edward.
Why this is different
Built around transfer, not content consumption.
The standard is simple: can you handle a problem you have not seen before with less guidance, clearer reasoning, and better technical decisions?
Not another course
No fixed sequence chosen before anyone has watched you work. The diagnostic determines what deserves attention.
Not framework collecting
Tools change. The mentorship develops the engineering model that lets you evaluate and use new tools well.
Not interview theater
Mock interviews become a test of capability already built—not a substitute for the underlying engineering work.
The decision today
You can start without deciding everything.
Reserve one diagnostic. Get a serious second opinion and a development plan. Decide on long-term mentorship only after you have seen the quality of the session and the evidence it produces.
- Live 1:1 with Edward
- 90 focused minutes
- Evidence-based capability map
- Prioritized first-phase plan
- No long-term commitment required
Questions worth answering
Before you decide
Will this treat me like a beginner?
No. Your existing experience is part of the evidence. The point is to locate the next constraint precisely, not erase what you already know or force you through introductory material you do not need.
Is this a job interview?
No. You are encouraged to think aloud, ask clarifying questions, state uncertainty, and explain how you would find an answer. The purpose is diagnosis and development, not pass or fail.
Why can’t I use AI for the whole session?
The diagnostic needs a clean baseline of your own mental model first. AI is then introduced intentionally so Edward can evaluate whether it improves your reasoning or replaces it.
Do I have to commit to long-term mentorship?
No. The diagnostic stands on its own. Any continuing engagement is recommended and scoped afterward based on the evidence, your goals, and the pace that makes sense.
What should I prepare?
Come ready to think aloud and work through unfamiliar material. You may bring questions about your development goals, but you do not need to study a hidden answer key or memorize a system-design template.