CBSE Class 10 · Artificial Intelligence, code 417 · Part C: practical and project work

The other fifty marks: the practical file, the project and both vivas, built across a year.

Half of the Class 10 AI subject is never written in an exam hall. It is a file of at least fifteen programs, a practical examination on the four practical units, a project that has to relate to the Sustainable Development Goals, and two vivas in which a student explains what they made and why. Schools conduct all of it before the theory paper, which means the family that treats it as a January job has already lost the January weeks that the theory paper needed. This page takes Part C component by component: what each is marked on, what a good file entry looks like, how a project gets chosen and documented, what the viva actually asks, and the month-by-month plan our batch runs so that none of it is a rush.

Source: CBSE curriculum 2026-27, AI (417) Class X, Part C · file built one program a week in the Wednesday lab

Start here

The courses that build the practical half

The 417 batch for the file and project, the tool CBSE names for the no-code units, and the course for students who want their project to be the best in the school.

Ans. The short version

Part C of CBSE Class 10 AI (417) is worth 50 marks and is conducted by the school before the theory paper: a practical file with a minimum of 15 programs (15 marks), a practical examination on Units 4 to 7, Statistical Data, Computer Vision, NLP and Advance Python (15), a viva (5), a project, field visit or student portfolio related to the Sustainable Development Goals (10), and a viva on that project (5). The file is built best at one program a week from the first month, the project is chosen in the first term and documented through the six stages of the AI project cycle, and both vivas are won by students who can explain their own choices. Modern Age Coders runs all of it inside the Monday and Wednesday 9 PM batch, closing the file and project by December.

Q1. What are the fifty marks made of?

Five components, and what each one is really judged on

ComponentMarksWhat the examiner is actually looking atWhere it goes wrong
Practical file, minimum 15 programs15Correct programs with their real outputs, indexed, in the school's format, that the student evidently wroteCopied files with outputs that do not match the code; programs the student cannot explain
Practical examination on Units 4 to 715An unseen task in Python or Orange, solved and shown running under timeStudents who rehearsed reproduction meet a variation and stop
Viva voce5Whether the student understands the programs in their own file"Sir, I wrote it but I do not remember why"
Project, field visit or portfolio10A documented AI project cycle on a problem linked to the Sustainable Development GoalsToo big to finish; no goal link; no evaluation numbers
Viva on the project5Whether the student made the decisions and can defend themA project built by an adult or a template, defended by nobody

Read the last two columns together and one thing stands out: every component is lost the same way, through work the student did not really do. A file copied from a senior, a project assembled from a template, a practical exam rehearsed as reproduction: each produces a student who cannot answer the follow-up question, and the follow-up question is where ten of the fifty marks live outright and the rest are quietly decided. The whole of this page is the alternative, which is slower and, over a year, easier.

Q2. What does a full-marks practical file look like?

Eighteen entries, one anatomy, built one Wednesday at a time

The minimum is fifteen. Eighteen gives the student a choice of programs to be asked about in the viva, and it costs three Wednesdays.

The anatomy of one entry

  • Serial number and date, matching the index at the front of the file.
  • Aim, one sentence, in the words CBSE's suggested list uses where it applies.
  • The program, exactly as it ran, with a comment line saying what each block does.
  • The output, the real one the program produced, which for charts and images means the actual picture.
  • A remark, one line: what was learned, or what would change with a different input.

The output rule catches more copied files than anything else: a program copied from a friend arrives with the friend's output, and an examiner who runs the program sees the mismatch in seconds. A file built in Jupyter, where output sits under code, cannot make this mistake.

Eighteen entries, by month

  • April to May: environment set-up, add two lists, mean, median and mode with NumPy.
  • June to July: line chart, scatter chart, bar chart, marks dictionary, CSV first ten rows.
  • August to September: CSV information, CSV filter, statistics report, display an image, image shape.
  • October to November: grayscale image, bag-of-words counter, train-test split by hand, accuracy calculator, menu program.
  • December: index, formatting to the school's rule, every output re-checked, file closed.

The eight CBSE suggested programs are written out in full, with their traps, on Python for CBSE Class 10 AI. The ten extra entries here are deliberately tied to theory units, so the file doubles as revision when February arrives.

Q3. How is the project chosen and documented?

One dataset, one question, one model, six documented stages, one goal

CBSE asks that the project relate to the Sustainable Development Goals and offers Palmer penguin species prediction as a sample. The winning shape is small, real and explainable.

  1. Problem scoping, and the goal chosen here, not at the end

    The 4Ws canvas from the project cycle: who is affected, what is the problem, where does it occur, why does it matter. The Sustainable Development Goal is named in this stage, because a goal bolted on to the last page is the most common viva wound. Waste photos to sort serve responsible consumption; rainfall and crop data serve zero hunger; sentiment on a city park's reviews serves sustainable cities.

  2. Data acquisition, with the source written down

    An open dataset with its origin named, or a small dataset the student gathered, which is the easiest thing in the world to defend. The project cycle asks for reliable sources and the examiner asks where the data came from; both are answered by one sentence in the report.

  3. Data exploration, with a chart the student can read aloud

    One or two charts, built in Orange or with Matplotlib, and a sentence each on what they show. This is the stage where the student discovers whether the question is answerable, and it is better discovered in September than in a viva.

  4. Modelling, in the tool that fits the student

    A classifier in Orange, following the penguins case study's shape, or a few lines of Python for students who prefer it. The model type is named and the choice is justified in one line using the Unit 2 vocabulary: supervised, classification, and why.

  5. Evaluation, with real numbers and an honest reading

    Accuracy at least, and for a classifier the confusion matrix with precision and recall. Then the sentence that earns the project viva mark: what the model gets wrong, and which metric matters for this goal. A project reporting 100 percent accuracy is a project the examiner will test.

  6. Reflection, and the report assembled

    What worked, what would change with more data, what the student would do next. The report is the six stages in order, a few pages, in the school's format. Scoped in September, reviewed monthly, finished in December: never a last-week build.

Q4. What does a viva actually sound like?

The same project, two vivas, ten marks apart

Both students built a classifier on a waste-sorting dataset. The examiner asks the same three questions.

The viva that loses markswork done, not owned
Q: Why did you choose this problem?
A: It was on a list. My cousin did it.

Q: Your accuracy is 92 percent.
   Is that good?
A: Yes, it is high.

Q: What does the model get wrong?
A: Nothing, it is 92 percent.

# Three answers, zero decisions.
# The examiner now suspects the
# file too, and checks it.

Nothing here is dishonest, and nothing here is defensible. The student did tasks; the questions were about choices, and there were none to describe.

The viva that earns marksthe same project, owned
Q: Why did you choose this problem?
A: Our school bins mix plastic and
   paper. Responsible consumption is
   SDG 12, and a photo sorter helps.

Q: Your accuracy is 92 percent.
   Is that good?
A: Good, but the confusion matrix
   shows it confuses foil with paper,
   so recall on foil is only 70 percent.

Q: What does the model get wrong?
A: Shiny paper. I need more foil photos.
   That is my next step.

# Three answers, three decisions.

Same dataset, same tool, same 92 percent. This student reads the matrix, names a metric, and knows the model's weak spot. That is ten marks, and it is teachable.

The difference is not intelligence or confidence. The second student scoped the problem themselves, read the confusion matrix in the Evaluating Models unit until it meant something, and practised saying what they did out loud, every Wednesday, in one sentence. That last habit is the cheapest ten marks in the subject and the one our batch never skips.

Q5. How is the practical examination rehearsed?

Fifteen marks on an unseen task, and the monthly mock that removes the fear

What the practical exam is

A task set on Units 4 to 7, solved live and shown running. In practice that means a Python program in the family of the suggested list with a twist, or an Orange task from the Statistical Data, Computer Vision or NLP practicals: load a dataset, build a workflow, read a result. The examiner is watching for a student who can start from a blank cell or an empty canvas, not one who can find the right page in a file.

How the batch rehearses it

From the second term, one Wednesday a month is a timed mock: an unseen task, solved alone with the camera on, then reviewed line by line. The first mock is usually uncomfortable and the fourth is routine, which is the point. Students also keep a one-page sheet of the twists they have met, so that the exam's twist is one they have already seen a cousin of.

The dress rehearsal in December ties it together: file checked entry by entry, project demoed and questioned, a timed practical task, and a viva with a teacher the student has not met. Students walk into the school practical having already done the whole day once. The theory paper's own countdown is on board exam preparation, and it begins where this page ends.

Q6. Where does Part C go wrong?

Six mistakes we see in files and projects every year

Outputs that do not match the code

The signature of a copied file. An examiner who runs one program and gets a different output stops trusting the whole file. Jupyter, where output lives under code, makes this mistake impossible to make by accident.

The mode that does not exist

np.mode appears in a startling number of files and never ran. NumPy has no mode; SciPy or the statistics module supplies it, and a student who knows why has turned the trap into a viva answer.

A project scoped for a company

Predicting the stock market, curing a disease, a chatbot for everything. Ten marks reward a small, finished, evaluated project with a goal link, and the ambitious one is abandoned by November.

The goal added on the last page

The Sustainable Development Goal belongs in problem scoping, where the 4Ws are. A goal that appears only in the conclusion tells the examiner the project was built first and justified after.

Evaluation with no numbers

"The model works well" is not an evaluation. Accuracy, a confusion matrix, and a sentence on what the model gets wrong are what the Evaluating Models unit taught, and the project is where the examiner checks it landed.

Everything in January

The school practical sits before the theory paper, so a January file steals the weeks the 21-mark theory units needed. Closing Part C by December is the single biggest planning decision in the subject.

Q7. What does it cost?

Monthly fees, the same for every course we teach

The practical work is part of the batch, not an add-on. Billed monthly, no admission fee, stop at any month end, and the free demo is a practical session.

Group batch · Mon and Wed 9 PM

₹1,499

per month

  • Ten to fifteen students, one teacher all year
  • File built one program a week, checked
  • Project scoped early, reviewed monthly, viva rehearsed
  • Certificate on completion
Book the free demo

Mini batch

₹2,999

per month

  • Four to five students
  • Other timings than the Monday and Wednesday batch
  • Every file entry and project stage reviewed individually
Ask about timings

One to one

₹4,999

per month

  • Private teaching on your own schedule
  • A behind file recovered in focused weeks
  • The honest choice from December onward
Enquire

What families say

Rated 4.9 across 547 Google reviews

Real reviews from real families. We neither write nor commission them.

★★★★★

"The one step solution for my son. Modern Age Coders make learning coding so simple that kids love it. The teachers explain complex concepts clearly with practical exercises and interactive content."

Ria Mukherjee

Parent

★★★★★

"Modern Age Coders has been a game-changer for me. I struggled to grasp IT concepts and coding before joining, but their classes transformed everything. I can now confidently write complex programs with ease."

Samriddha Mondal

Student

★★★★★

"One of the most wonderful education centres out there. Education is not limited to school syllabus but focuses on skill development."

Vansh Agarwal

Student

★★★★★

"My child Dhairya is really enjoying the Modern Age Coders classes. This is his first online class and he eagerly looks forward to it. I can already see his improvement, and the teachers are very cooperative."

Sonam Oswal

Parent of Dhairya

★★★★★

"Modern Age Coders have wonderful teachers who teach in a clear, easy and practical way. The teacher boosts students' confidence and inspires them to learn without hesitation."

Sonu Goyal

Parent

★★★★★

"I highly recommend this computer coding class! The teachers are incredibly knowledgeable and passionate about coding."

Ritu Kedia

Parent

The rest of the Class 10 series

Nine more pages for Class 10 board students

Four more on CBSE AI 417, and five on the ICSE Computer Applications paper for families on the other board.

Questions about Part C

What families ask about the file, the project and the vivas

Who conducts the 50-mark practical for CBSE Class 10 AI?

The school, in the weeks before the theory paper, following CBSE's Part C scheme: 15 marks for the practical file, 15 for a practical examination on Units 4 to 7, five for a viva, ten for the project, field visit or portfolio, and five for a viva on that project. The school announces its own dates, and because they fall before the theory exam, everything on this page needs to be finished by the end of December in a normal year.

Can the project be built in Orange rather than in Python?

Yes. CBSE's own sample project, predicting Palmer penguin species from statistical data, is the Orange Data Mining case study from the Statistical Data unit, and the curriculum names Orange, Teachable Machine and Lobe as tools for the practical units. A project in Orange is fully legitimate; what earns the marks is the documented project cycle and the student's ability to explain each stage, not the tool. Students comfortable in Python may prefer it, and either is fine.

What are the field visit and portfolio alternatives to a project?

CBSE lets the ten marks come from any one of project work, a field visit or a student portfolio. A field visit means visiting an organisation that uses AI and reporting on it; a portfolio is a curated collection of the student's AI work across the year. Most students choose the project because it is the easiest to control from home and the easiest to defend in the viva, and our batch prepares that route by default while helping any student who has a genuine field-visit opportunity.

Does the project have to relate to the Sustainable Development Goals?

The curriculum asks that the project, field visit or portfolio relate to the Sustainable Development Goals, so yes, and the link should be explicit in the report's problem scoping rather than a line added at the end. It is not a constraint in practice: a classifier that sorts waste photos serves responsible consumption, a dataset on rainfall and crop yield serves zero hunger, a sentiment model on a city's park reviews serves sustainable cities. The scoping stage of the project cycle is exactly where the goal is chosen.

How long should the project report be?

As long as the six stages of the project cycle need and no longer: problem scoping with the goal and the 4Ws, data acquisition with the source named, data exploration with a chart or two, modelling with the tool and the model chosen, evaluation with real numbers, and a reflection on what the model gets wrong. Schools set their own formatting rules, but a report that documents those stages honestly in a few pages beats a thick report that pads them. The viva examiner is looking for the cycle.

Should the practical file be handwritten or printed?

That is the school's decision and it varies, so ask early and follow it exactly. What does not vary is the anatomy of a good entry: a serial number, the aim, the program, the output actually produced, and a one-line remark, all indexed at the front. We build the file in Jupyter first, where the program and its output sit together, and then produce whatever physical form the school asks for from that source.

What does the viva examiner actually ask?

For the file viva, questions about specific programs: what a line does, what would change if an input changed, why a library was used. For the project viva, questions about decisions: why this goal, why this data, why this model, what the evaluation number means and what the model gets wrong. Both vivas reward a student who made the choices and remember why, which is the entire reason our batch has students explain every program aloud in one sentence every week.

Do you provide datasets for the project?

We help students find open datasets that fit the goal they chose, and we teach them to say where the data came from in the report, which is a scoping-stage requirement. Students who collect their own small dataset, photos of classroom waste or a survey of their building, are encouraged, because a dataset the student gathered is the easiest thing in the world to defend in a viva.

What does the practical preparation cost?

The practical work is part of the standard batch, not an add-on: the group batch, the mini batch and one to one are shown on this page in your own currency, billed monthly with no admission fee. The free demo class is a Wednesday-style practical session, so a family sees exactly how the file gets built before deciding anything.

My child's file is behind. Can it be recovered?

Almost always, if the recovery starts before November. A student with a real understanding of the eight suggested programs can write the whole file in a few focused weeks with someone checking each entry, and the project can be scoped to something small enough to finish. From December, one to one is the honest route, because the school practical is close and the theory revision needs the January weeks. The free demo tells you which situation you are in.

Start

The free demo writes the first file entry

Bring whatever file exists, even if it is empty, and whatever project idea exists, even if it is "something with AI". In the demo the student writes one file entry properly, aim, program, real output, remark, and the teacher runs the three scoping questions on the project idea until it is one dataset, one question and one goal. You leave with a file that has started and a project that is possible, and a clear read on how many Wednesdays stand between your child and a closed Part C.

Reading further first? Python for Class 10 AI writes out the eight suggested programs, and the syllabus explained shows where Part C sits in the whole.

WhatsApp us · +91 91233 66161 · contact@modernagecoders.com

Every class is live video. The form books one callback and holds no seat against your name.

We call within a few hours. No card, no enrolment fee.

Keep exploring Modern Age Coders

By class and age

Learn more

Free resources

From the blog

Start here