Jiwaji University Students: A Mid-Internship Review Template to Self-Assess and Reset Goals

The best time to review a 14-day AI internship is the end of day 7, and the best way is a written form: rate each goal, name what is blocking you, and rewrite the second week around what is realistically finishable. The template below does that in about 40 minutes and is built so you can send it to a supervisor without editing.

Two weeks leaves little room for drift. If you only notice on day 12 that your model was never evaluated properly or your notebook is unreadable, there is nothing left to fix. A midpoint review turns vague worry into a short list of decisions.

Why a day-7 review suits a 14-day AI internship

Longer internships can absorb a bad week. A 14-day one cannot. By day 7 you have normally finished setup, read the brief, explored the data and perhaps trained a first baseline. What remains is the part that produces something you can show: evaluation, improvement, documentation and a final summary.

AI work has some habits that make a review especially useful:

• Scope creep through tooling. Students spend days comparing libraries or models when a simple baseline would do.
• Invisible progress. Hours of data cleaning leave nothing visible unless you write down what changed and why.
• Unmeasured results. A model that “seems to work” has no value until you pick a metric and a validation split.
• Environment problems. Package conflicts, missing GPU access and file-path errors can eat a whole day.

If you are doing the free two-week online internship from CareerSwami, the same logic applies: you receive an offer letter at the start and a completion certificate after successful completion, so what you want by day 14 is finished, explainable work rather than a pile of half-started experiments.

The midpoint review form: sections and what each one asks

Copy the form into a document or notebook. Fill it in honestly, because a supervisor can only help with problems you state. It has five parts: snapshot, progress ratings, blocker log, revised plan, and requests for your supervisor.

Part 1: Snapshot (5 minutes)

• Name, programme and semester or year, track chosen, internship dates.
• Original goal in one sentence (copy it from your day-1 plan).
• Hours spent so far, roughly, and on which days you could not work.
• One-line status: on track, slightly behind, or significantly behind.

Part 2: Progress rating scorecard (15 minutes)

Rate each area from 0 to 4, using the descriptors in the table. Add one line of evidence for every score, such as a file name, a commit, a chart or a note. A score with no evidence should drop by one point.

Add your scores and divide by 28 to get a rough percentage. Treat it as a prompt, not a grade. A 40% score on day 7 with a clear plan is healthier than 70% with unexplained work.

Part 3: Blocker log (10 minutes)

Write each blocker as a row. Be specific: “TensorFlow import fails with a version error” is useful; “coding is hard” is not.

Sort the log by days lost. Anything above one day lost needs either an outside helper or a decision to change scope.

Part 4: Revised plan for days 8 to 14 (10 minutes)

Rewrite the second week with fixed outputs. Use the table below as a skeleton.

Then state three revised goals using the must, should, could pattern:

• Must: the minimum deliverable that makes the internship worth showing.
• Should: the improvement you will attempt if the must is secure by day 11.
• Could: an extension you will only touch if everything else is done.

Part 5: Requests for your supervisor (5 minutes)

• One decision you need from them (for example, “Is accuracy or F1 the right metric here?”).
• One resource you need (a dataset, a reference, a clarification).
• One thing you want feedback on before day 14.

Worked example: a day-7 review for a sentiment classification task

This example is illustrative. The numbers are invented to show how the form is filled in; no code was run.

A student is building a sentiment classifier for product reviews. At day 7 she fills in the scorecard:

Total: 14 of 28, so 50%. Her blocker log shows one significant row: “Notebook crashes on the full dataset, 1.5 days lost, tried reducing batch size, supervisor can advise on sampling, resolve by day 8.”

Her insight is that accuracy looks fine only because one class dominates. Her revised plan therefore changes the metric to macro F1, fixes a stratified split on day 8, drops her planned comparison of three deep models, and keeps a single logistic regression baseline plus one improved version. Her must goal is a documented, evaluated baseline; her should goal is a class-weighting experiment; her could goal is a small error analysis chart. The request to her supervisor is: “Is macro F1 acceptable as the main metric?”

Notice that the review did not add work. It replaced ambitious but unmeasured work with smaller, verified work.

How to turn your total into a decision for week two

• 22 to 28 (on track): protect the plan, add one “should” goal and begin the README early.
• 14 to 21 (recoverable): cut at least one planned item, fix the lowest-scoring area first and send a written update to your supervisor today.
• Below 14 (behind): reduce to a single must goal, such as one dataset, one baseline, one metric and one write-up, and ask for a short call to agree scope.

Also check for lopsided scores. A high score in model building alongside zeros in documentation and evaluation usually signals that you are coding without producing anything reviewable.

Sharing the review with a supervisor without it feeling like a complaint

Keep the message short and put the decision first. A workable structure:

1. Subject line: “Day 7 review: [your name], [track]”.
2. Two sentences summarising status and the main blocker.
3. The completed form attached or pasted below.
4. One clear question and a suggested time to discuss.

Share your evidence links with the form, since supervisors respond faster to a notebook link than to a description. If you do not have a named supervisor, use the same form as a self-contract: save it, then compare it with a day-14 version.

Jiwaji University students: fitting an AI internship review around your programme

Jiwaji University is a state university in Gwalior. Its official homepage states it was established on 23 May 1964, displays a NAAC A++ grade and counts about 420 colleges, so many students study at an affiliated college rather than a university department. That matters for this template because internship rules, where they exist, may be set at either level, and a short online internship has to fit within your own timetable.

Programme documents are the next place to look. The university lists NEP-era syllabus links for MCA 2025-27, MBA E-commerce 2025-27 and BCA 2025-29, and a separate page on NEP syllabus under Ordinance 14(1) and 14(2) carries a Hindi notice whose title refers to annual and semester systems. Check the current version of these pages yourself, because the details of how your course is structured affect how many free days you will have. The School of Studies in Computer Science page carries e-learning notes for BCA, MCA, MSc (CS) and related courses. These are teaching materials rather than a scheme document, but they may help you find a concept to revise when your blocker log shows a knowledge gap.

The homepage also links a Career Counseling Cell and Placement Cell. Whether either cell has a view on short online internships is not something this article can state, so ask them directly.

No official internship, industrial training or project-work rule for any programme was verified for this article. Do not assume that this internship carries credit, or that your college will recognise it, until you have it in writing.

Four to six verification steps for Jiwaji University students

1. Open the syllabus or scheme PDF for your exact programme and entry year, and look for any internship, training or project component, its semester and its evaluation method.
2. Ask your department head or class coordinator in writing whether a voluntary online internship needs permission, and whether the certificate can be added to your records.
3. Find out if your college follows the annual or semester pattern for your batch, and mark your exam and practical dates before planning week two.
4. Email or visit the Career Counseling Cell or Placement Cell and ask whether they accept self-found internship certificates, and what format they prefer.
5. Save any reply as a PDF or screenshot, so your supervisor review and your college file tell the same story.

Adapting the form to common AI internship tasks

The scorecard is generic by design. Swap the evidence examples to suit your task:

• Classification or regression: evidence is a validation metric, a baseline for comparison and a short error analysis.
• NLP: evidence is a tokenisation choice you can justify, a sample of misclassified text and a note on dataset bias.
• Computer vision: evidence is an augmentation record, a confusion matrix and a few sample failure images.
• Prompt or LLM-based tasks: evidence is a set of test prompts, a pass/fail record and notes on inconsistent outputs.
• Data analysis and visualisation: evidence is a cleaned dataset description, charts with titles and a written insight for each.

Day-14 follow-up: closing the loop on your day-7 review

On the last day, reopen the day-7 form and add a column beside each score. Mark where you improved, where you did not, and why. Then write three sentences: what changed in your plan, what you would do differently in the first week, and which skill you want to practise next. That short note makes a strong reflection for a portfolio, a project README or an interview answer, because it shows judgement rather than only activity.

Frequently Asked Questions

When exactly should I do the mid-internship review in a 14-day programme?

Do it at the end of day 7, or the morning of day 8 at the latest. Earlier than day 6 you will have little to rate, and later than day 9 you cannot change much. Block 40 minutes and finish before starting new work.

How do I rate myself fairly if I am a beginner in AI?

Score against the task, not against experienced practitioners. The descriptors in the scorecard are about observable outputs, such as a stated metric or a fixed split, so a beginner can score well on process even if the model is simple. Always cite one piece of evidence per score.

What if my supervisor does not reply to the review?

Send a short follow-up after a day with your single most important question. Meanwhile, continue with the must goal of your revised plan so you are not waiting. Keep a record of what you sent and when.

Should I cut scope or work longer hours when I am behind?

Cut scope first. Longer hours often produce more unmeasured experiments. Fix the biggest blocker, reduce to one dataset and one baseline, and spend saved time on evaluation and documentation.

Can I use the review form for an internship that is not AI?

Yes. Keep the structure of ratings, blockers, plan and requests, and replace the AI-specific rows with your own areas, such as design, testing or content work. The evidence rule still applies.

As a Jiwaji University student, how do I find out whether my college will recognise this internship?

No internship rule was verified for this article, so ask. Write to your department head or placement coordinator, name your programme and entry year, and ask whether voluntary online internships are recognised and in what form. Keep the reply on file.

How long should the revised plan be?

One page is plenty. Seven days, one output per block, three goals and a list of things you will cut. If it takes more than a page, the scope is still too large.

Does the review need to be shared at all, or is a private version enough?

A private version helps you think, but sharing it adds accountability and gets you faster answers on blockers. Share at least the blocker log and your question.

Related Artificial Intelligence guides

• University of Kota Students: Preparing for AI Internship Coding Screens with a Practice Plan
• F.S. University Students: Writing a Model Card for Your AI Internship Project
• ASBM University Students: AI for Business Analytics Internships, From Dashboard to Prediction
• All Artificial Intelligence internship guides