IB DP Math AI HL Online Tutor in Dubai

Four weeks of weekly AI HL support should leave four dated artefacts behind. Week one produces a written diagnosis built on an existing script. Week two produces the first piece of work set and marked. Week three produces the first repeat test of the original weakness. Week four produces a short pattern note saying whether the weakness has moved. If any of those four is missing, the question to ask is not whether your child is enjoying the sessions.

The trail matters because mathematics improvement is not linear and feels like nothing for the first fortnight. A student rebuilding how they set up a model will temporarily get slower and may score lower, and without dated evidence that looks like failure rather than reconstruction. With the scripts in front of you, it is visible as a method change: fewer marks on the first attempt, the same marks regained two weeks later, and then marks appearing that were never there before.

What should be in each note is specific and short. The topic, the component it belongs to, the marks available, the marks scored, the mark type lost, and the single thing set as follow-up. Notes longer than that tend to be padding, and notes shorter than that are not evidence. We write them in plain language rather than in assessment vocabulary, because a note a parent cannot read is a note that gets filed and never used.

Missing evidence has honest explanations. A session lost to a school trip, a week where the student had three internal deadlines, or a deliberate decision to spend two sessions teaching rather than testing are all legitimate. What is not legitimate is four weeks of pleasant sessions with nothing written down, and a family being asked to renew on the strength of a feeling. The table sets out what to expect and what to do if it is absent.

Paper 3 is one hour, 55 marks and 20% of the AI HL grade, made up of two compulsory extended-response problem-solving questions. Technology is required throughout. Its structure is what makes it different: each question unfolds across several parts that depend on earlier answers, so a modelling choice made in part (b) can cost marks in parts (d), (e) and (f) even when every subsequent step is executed correctly. No other component in AI HL punishes an early decision that way.

The practical consequence for Dubai families is that Paper 3 is under-evidenced. School mocks often use a shortened version or omit it, because marking two long investigations for a whole cohort is expensive in teacher time. A student can therefore arrive in February of DP2 having never completed a full-length Paper 3 under timed conditions. The first full attempt is almost always a shock, and it is far better experienced in a session in November than in an examination hall in May.

A marked Paper 3 script is also the most informative single document a tutor can produce about an AI HL student. It shows whether they can read a long stem, whether they define variables before using them, whether they notice that part (e) is asking them to validate the model they built in part (b), and whether they write a conclusion in the context of the question rather than as a bare number. Those four habits are worth more than any amount of topic revision.

Ask for the attempt, the markscheme-aligned marking and the propagation trace: which later marks were lost because of which earlier decision. A tutor who can produce that for one paper can produce it for the rest. One who returns a percentage and a verbal summary has not done the work that makes the evidence useful.

A mark on an IB mathematics markscheme is not a generic point. Method marks reward a correct approach even when the arithmetic fails afterwards. Accuracy marks reward a correct value, usually to three significant figures unless the question says otherwise. Reasoning marks reward a justification written in words. Follow-through allows a correct method applied to an earlier wrong answer to still earn credit. A student who does not know which type they are losing cannot fix the loss, and neither can a tutor working from a percentage.

This distinction changes what a session should do. A student losing method marks needs to be taught a technique. A student losing accuracy marks while keeping every method mark does not need teaching at all; they need discipline around rounding, units and transcription, and that is fixed by habit rather than by explanation. Buying teaching hours for an accuracy problem is one of the commonest wasted spends in the Dubai tutoring market, and the marked script is what exposes it.

Follow-through is the rule that most changes how a parent should read a low mark. A student can make one error in part (a) and still collect most of the marks in later parts, provided their working is visible. The corollary is brutal: a student who writes down only final answers forfeits both the method marks and the follow-through. On Paper 3, where parts chain together, invisible working can turn a single slip into a double-figure loss.

We write the mark type beside each loss on anything we return, and the six-week pattern note counts them. Four weeks of accuracy losses and no method losses is a very different brief from the reverse, and the two briefs need different tutors, different work and a different number of hours. That is the whole argument for evidence over impression.

The AI HL exploration is marked out of 20 across five criteria: presentation, mathematical communication, personal engagement, reflection, and use of mathematics, with six of the twenty marks on the last of those. It carries the same 20% weight as Paper 3. The IB recommends a piece of roughly twelve to twenty pages. The evidence that it is genuinely the student's own is a dated version history, and that history is also the best protection either of you has if authorship is ever questioned.

Practically, that means keeping versions rather than overwriting one file. A first version with the topic and the data source. A second with the mathematics attempted and the gaps visible. A third after feedback, where the changes are the student's response to questions rather than text supplied to them. Dubai schools run their own internal deadlines and their own authenticity conversations, and a student who can walk a coordinator through three dated versions is in a different position from one who cannot.

The boundary on paid support is firm and we state it the same way everywhere. A tutor may teach a technique on a different example, ask what a symbol means, say where a reader loses the thread, and ask what would change the result. The student makes every choice and types every word of what is submitted. Any provider offering a finished document for a fee is selling academic malpractice, and the consequence under IB rules falls on the candidate, not the seller.

At Higher Level the commonest failure is not dishonesty but ambition. A student picks a topic requiring mathematics beyond the course, spends six weeks on it and produces something that scores poorly against criterion E because the mathematics is unsupported rather than wrong. One session testing topic viability against the AI HL toolkit in the first fortnight prevents most of that, and it produces a written note recording what was ruled out and why.

A grade in IB mathematics is produced by adding weighted component scores and comparing the total against boundaries set for that examination session. The boundaries are not fixed in advance and are not ours to publish; grade-boundary tables are the IB's material, and a tutoring page reproducing them would be both a copyright problem and a disservice, because a boundary from one session does not transfer cleanly to the next. What a family can work with instead is raw marks and the weighting arithmetic.

That arithmetic is worth doing once in front of the scripts. A student scoring consistently on Paper 1 and Paper 2 but leaving Paper 3 largely unattempted is carrying a 20% component at a fraction of its value, and recovering even half of that component moves the total more than a general improvement across the two long papers would. Conversely, a student losing a scatter of accuracy marks across all three papers has no single fix, and the honest answer is that the gain will be smaller and slower.

This is also where we decline to make a promise. Nobody can commit to a particular grade, because the total depends on components we do not mark and boundaries nobody knows in advance. Any provider promising a particular score in advance is describing something they do not control. What a tutor can commit to is the process: marked work every week, a count of mark types lost, and an honest statement at six weeks about whether the pattern is moving.

What progress looks like in the numbers, then, is this: the same mark type appearing less often across successive dated scripts, and a previously unattempted component starting to score. Those two movements are visible weeks before a school report reflects them, which is exactly why the evidence trail is worth more to a parent than a predicted grade.

Statistics and probability takes 52 of the 240 guide hours at AI HL, more than any other topic, with geometry and trigonometry at 46, functions at 42, calculus at 41 and number and algebra at 29. The weighting surprises students who chose AI expecting a lighter course than Analysis and Approaches. It is not lighter; it is differently heavy, and the weight sits on interpretation of data rather than on manipulation of symbols.

In the marked scripts, the losses concentrate in identifiable places. Choosing the wrong statistical test or regression model for the data given. Quoting a correlation coefficient without saying what it means for the situation described. Treating a p-value as a verdict rather than as evidence. Confusing the conditions under which a normal model applies. Each of those produces a distinctive pattern of lost reasoning and context marks, and each is fixable in two or three sessions once named.

The Higher Level additions are where Paper 3 draws much of its material: more demanding probability, matrices and graph theory applied to real contexts, and modelling that has to be validated rather than merely constructed. A student who can build a model but never checks it against the data will lose marks consistently in the back half of a Paper 3 question, and that is the single most common HL-specific pattern we see in Dubai scripts.

The practical sequence is to test statistics first in any diagnostic session, because it is the largest block and the one most likely to be generating losses. If the diagnostic shows statistics is secure and calculus is not, the plan changes accordingly. The point of leading with evidence rather than a standard syllabus tour is precisely so that the plan can be wrong in week one and corrected in week two.

Lessons for UAE families run one-to-one over live video, on a shared screen where tutor and student write on the same page, timed to Gulf Standard Time around a Dubai school day. IB Gram does not offer home visits in Dubai. Home visits are a service we run only in India, and no sentence on this site should suggest otherwise to a family in this emirate.

The case for that is about the evidence rather than convenience. The International Baccalaureate accounts for 7 per cent of Dubai private school enrolment against 37 per cent for UK curricula, which means AI HL specialists are thin on the ground in any single community. Requiring a tutor to be within a tolerable drive of your villa reduces the pool to whoever is nearby and available, and the person who is nearby is often a general maths tutor rather than someone who has marked Paper 3 scripts to a markscheme.

Online delivery also improves the artefacts themselves. Work is shared as a file rather than carried in a bag, so a marked script exists in two places and can be read by a parent in another country. The shared screen leaves a written record of the session's working. Nothing is lost to a dropped folder or a forgotten exercise book, and a tutor who moves house or changes city does not end the arrangement mid-course.

The honest counterweight is that a student who will not sit still for a video lesson is not served by one, and we say so at the trial rather than afterwards. The trial exists for exactly that question. The comparison table sets the options against each other on the terms that matter for evidence: marking, turnaround, the pool you choose from and what survives if the tutor changes.

Everything described on this page creates records about a child: marked scripts, session notes, a school name, a year group and a set of grades. UAE Federal Decree-Law No. 45 of 2021 on Personal Data Protection governs that, and the honest position is to tell you what exists rather than bury it in a policy page. We keep the session notes, the marked work and the enquiry details. We do not collect a child's photograph, their full school timetable or their medical information, because none of those are needed to teach mathematics.

Deletion is a request, not a negotiation. Ask and the file goes, including the marked scripts, and we confirm when it has. The one thing we will do first is offer to send you copies, because a family that has paid for a term of marking usually wants the scripts even after they stop the lessons. Several providers in this market treat accumulated student work as their own asset; it is the student's.

The other half of honest reporting is what we refuse to publish. No satisfaction percentage, no median tutor score, no claim about what proportion of students improved, because we cannot show you the sample, the date or the method behind any of those numbers. Pages in this search carry all three, and in some cases two contradictory versions on the same page. A statistic without its sample size and its method is decoration.

What a session note will say is narrow and checkable: the date, the component, the topic, the marks available, the marks scored, the mark types lost and the single piece of follow-up work. What it will never say is that your child is on track for a particular grade. That claim depends on components nobody in this arrangement marks, and a tutor who makes it is telling you something they cannot know.