The instruction was consistently adapted to the learner's current level and preferred way of learning.
Artificial intelligence learnerVerify on SuperProf
Project-Based AI Mentorship
AI Project Mentorship
AI project mentorship at Pristone Academy is a bounded live 1:1 engagement for students who know some Python and are ready to build something more substantial than a tutorial. The student defines the problem, designs the system, implements and tests it, evaluates the result, and remains able to explain and continue the work independently.
Premium private coaching, typically structured as multi-session engagements.
Project mentorship is premium live 1:1 work scoped after a paid diagnostic. The diagnostic defines a bounded milestone plan, likely session cadence, and exact rate before scheduling. Research-length work is moved into the separate Research Mentorship Track rather than allowed to expand invisibly.
Replies within 2 hours.
The Pristone Method
One capability loop across every track.
- Diagnose
- Prioritize
- Build
- Test transfer
- Measure
- Re-prioritize
Who this is for
Students who have completed introductory coding, can write and debug basic Python with some independence, and want to build a complete AI, data, simulation, or machine-learning project before deciding whether formal research is the right next step.
- Students moving from coding exercises to a complete technical system
- Learners who have an idea but need help reducing it to a feasible first version
- Students preparing for more advanced machine learning or research work
- Builders who want architecture, testing, evaluation, and code-review feedback
What you walk away with
- ✓A scoped project brief with an explicit user, question, or technical objective
- ✓A coherent architecture and working end-to-end implementation
- ✓Testing or evaluation tied to the behavior the project is supposed to demonstrate
- ✓Documentation of decisions, limitations, AI assistance, and credible next steps
Who this is not for
Being upfront about fit saves everyone time — including yours. This program is not a good match if:
- Anyone seeking a ready-made portfolio artifact or completed school assignment
- Projects that require confidential, private, or improperly sourced data
- Students who cannot yet write and debug basic Python without constant line-by-line direction
- Families seeking guaranteed admissions or competition outcomes from a project
Relevant verified feedback
Read the teaching record before you book.
These are privacy-safe paraphrases of feedback published on independent tutoring marketplaces. No student names or identifying details are reproduced here.
The teaching connected the big picture to the details and stayed attentive to what the learner actually needed.
Adult technical learnerVerify on Wyzant
These reviews support Edward's adaptive technical instruction. They do not claim that a particular project or admissions outcome resulted.
From idea to defensible project
1
Readiness and scope
Review the student's current code, define the user or question, identify prerequisites, and reduce the idea to the smallest meaningful end-to-end version.
2
Architecture and baseline
Choose the system boundary, data flow, components, interfaces, and a simple baseline that makes later complexity easier to evaluate.
3
Build and review
Implement in milestones, use code review to improve structure and debugging behavior, and keep the student responsible for every major decision.
4
Test and evaluate
Check important behaviors and failure cases; for model-based work, use an appropriate split, metric, and comparison rather than relying on a compelling demo.
5
Explain and continue
Document the architecture, evidence, limitations, and next milestone so the student can present and extend the work without the mentor.
Diagnostic map
What the AI project diagnostic examines
Python independence
How the student reads, writes, decomposes, and debugs basic programs without line-by-line prompting.
Problem definition
Whether the idea has a clear user, decision, prediction target, input, output, and success condition.
Architecture readiness
How the student breaks a system into components, manages data, and reasons about interfaces and failure.
Evaluation judgment
Whether the project can be tested honestly and whether a simpler baseline should come first.
The first milestone should be smaller than the ambition.
A project diagnostic turns a broad idea into a feasible system boundary, a baseline, and a sequence of work the student can own.
Frequently asked questions
A project must solve or demonstrate something coherently. Research adds a defensible question, explicit methodology, stronger evidence requirements, and a formal account of limitations. A project can be valuable without claiming novelty.
A complete beginner will usually benefit from coding foundations first. The diagnostic can recommend a preparation block and define the observable skills needed before the project starts.
Yes, when its use is appropriate, disclosed, tested, and understood. Generated code or text does not replace the student's responsibility to explain and verify the work.
Only with the required written permission and a privacy review. The student keeps ownership. Public presentation is optional and is not required for the project to be educationally successful.
Possible directions include recommendation, forecasting, classification, computer vision, language applications, simulation, optimization, or AI-enabled software. Feasibility and data quality determine the final scope.
Before you book
Clear answers and clear scheduling policies.
Ask first, free
Send Edward a question before paying or scheduling. Replies within 2 hours.
Rescheduling
Reschedule with at least 24 hours' notice at no charge.
Cancellation and no-shows
Cancel with at least 24 hours' notice. Cancellations inside 24 hours and no-shows count as used unless an emergency exception is agreed.
Satisfaction and outcomes
No score, admission, job, or other outcome is guaranteed. If a session materially differs from its published scope, contact Pristone within 48 hours so it can be reviewed and an appropriate correction proposed.
Explore next
Move from tutorials to a complete, student-owned project.
Begin with a project diagnostic or ask Edward about prerequisites, scope, data, and whether the idea belongs in project mentorship or the research track.
Founder-led by Edward Mabonga (MIT EECS) · 250+ five-star reviews · 1,000+ tutoring hoursView verified reviews