Learn the work behind Data, AI & Forward Deployed Engineering
Practical explanations, career decisions and reproducible workflows. Read the reasoning, inspect the evidence and follow the next skill into a real programme.
Evaluate recommendations with temporal holdouts
A random interaction split can train on a user's future and test on their past. Recommendation systems serve forward in time, so an offline split should reproduce what the system knew before each held-out interaction.
Evaluate retrieval on numerical tables
Table questions require the right row, column, unit and condition. A response can quote the correct number from the wrong band and still look plausible.
Evaluate retrieval recall at k with labelled evidence
Generation cannot cite evidence that retrieval never supplied. Recall@k measures whether at least one labelled relevant item appears among the first k eligible results for an answerable question.
Evaluate sentiment on mixed and sarcastic examples
A sentiment model can score clear praise and complaints well while failing exactly where interpretation matters. Mixed statements contain competing aspects. Sarcasm can reverse literal wording. Evaluate these as named sl
Excel data cleaning: preserve the raw sheet and document changes
Clean spreadsheet data should remain traceable to what arrived. Keep an unchanged source snapshot, transform a separate working table and record rules that explain every dropped row, changed label or rejected value. A po
Excel date serials and imported timestamps explained
Excel dates are numeric values displayed through date formatting. A timestamp's time-of-day component is represented by a fraction of a day. Text that looks like a date is a different kind of value and may behave differe
Excel formula errors: fix the cause before using IFERROR
IFERROR changes what a formula displays when it encounters an error. It does not correct the underlying calculation, restore missing data or distinguish every possible cause. Replacing errors with zero can turn a visibly
Excel missing values: distinguish blank cells and empty strings
Two cells can look empty while containing different things. One may be truly unused; another may contain a formula returning an empty string; a third may contain a space. Numeric zero is different again. Choose a missing
Excel percentage points versus percentage change
A rate rising from 20% to 25% increases by five percentage points and by 25% relative to its previous value. These are different descriptions of the same change. Label the unit explicitly so readers do not confuse an abs
Excel scenario analysis with explicit assumptions
Scenario analysis asks what a model produces under specified assumptions. It does not establish that those assumptions will occur. Keep inputs, calculations and outputs separate, and label hypothetical scenarios clearly
Excel SUMIFS with dates and inconsistent category labels
SUMIFS adds values that meet several conditions, but it cannot repair incorrect dates, duplicated rows or inconsistent category mappings. Establish a clean reporting table first, then express each eligibility rule explic
Excel weighted averages for price and discount analysis
A weighted average gives each observation influence according to a meaningful denominator, such as units sold or eligible gross value. An ordinary average gives each row equal influence, which can be misleading when rows
Not sure which programme fits?
Tell us your background and we will map it to the right entry point — including saying so when a cheaper programme is the better fit. A counsellor replies within one working day.