JNU Students from Humanities and Social Sciences: Entering an AI Internship Without a CS Degree

Yes, a JNU student from history, sociology, political science, economics, linguistics or a language programme can enter an AI internship without a CS degree. Entry-level AI work rewards people who can read carefully, frame questions, handle messy text and explain results, as long as they add basic Python and a few data habits. This guide gives you a skills-gap map, a 14-day project pathway built around text analysis of policy documents, and a worksheet that turns your training into something an AI team can use.

What an entry-level AI internship asks of you, and what it does not

People outside the field often picture AI work as advanced mathematics and building neural networks from scratch. A short internship rarely looks like that. Typical beginner tasks include cleaning a dataset, labelling text, running an existing model, checking its outputs for errors, writing up what was found, and presenting it to non-technical people. Several of those depend on judgement about language, context and bias, which is where humanities training is strong.

What you do need is a working floor of technical skill. You should be able to read a CSV file, run a short Python script without panic, and describe in plain words what a model did and where it failed. You do not need a degree to reach that floor in a few weeks of focused effort. You need a plan, which is what follows.

Skills-gap map for humanities and social science students entering AI

Use this table to place yourself honestly. The middle column shows what your programme probably already trains. The right column shows the gap and a realistic first step. These are general patterns, not claims about any specific JNU course.

How to read your gaps

• Mark each row as “comfortable”, “partial” or “not yet”.
• Pick no more than two “not yet” rows to work on in a 14-day window. Python basics and text data handling are the usual best choices.
• Treat rows where you are strong as the part of your pitch that you lead with.

Why policy documents make a good first AI project for a humanities student

Policy documents are public, text-heavy and full of structure: committee reports, white papers, parliamentary questions, scheme guidelines, budget speeches, court summaries. You already know how to read them critically. That matters because the hardest part of a text-analysis project is deciding whether the output makes sense, and a computer cannot do that for you.

A small project also gives you something concrete to show. A reviewer can see a question, a method, a result and your interpretation, which says more about your readiness than a list of courses.

A 14-day text analysis pathway using policy documents

The plan below fits a free two-week online internship such as the free two-week internship on CareerSwami, which includes an offer letter and a completion certificate after successful completion. Treat it as a pacing guide you can adapt to your assigned track, not a promise about what any internship will assign.

Days 1 to 3: question, corpus and setup

• Choose one narrow question, for example: how does the emphasis on a theme change across several years of the same type of document?
• Collect 8 to 15 public documents from official government or institutional websites. Note each source and date in a sheet.
• Set up a notebook environment and confirm you can open a text file and print its first 200 characters.

Days 4 to 6: cleaning and counting

• Convert documents to plain text and remove headers, page numbers and tables that distort counts.
• Count the most frequent meaningful words after removing common function words.
• Write down two things the counts suggest and one reason they might be misleading.

Days 7 to 9: a simple method

• Pick one technique: keyword-in-context search, a basic sentiment or tone check, or grouping paragraphs by similarity.
• Hand-label 30 to 50 paragraphs yourself first, so you have your own reference to compare the tool against.
• Run the method and compare its output with your labels.

Days 10 to 12: evaluate and critique

• Calculate how often the method agreed with you and look closely at the disagreements.
• Note patterns in the errors, such as sarcasm, legal phrasing, Hindi-English mixing or technical terms.
• Draw one clear chart that supports your main finding.

Days 13 to 14: write up and package

• Write a one- to two-page note: question, data, method, result, limits, next step.
• Prepare a short explanation you could give aloud in two minutes.
• Store the notebook, the source sheet and the note together so someone else could follow your steps.

Illustrative walk-through: tracking how a policy theme is framed

This example is illustrative. It describes a plan, not results from code that was run.

Suppose a student of political science wants to study how a set of public reports talks about “skilling”. She gathers ten reports from different years and extracts the text. After cleaning, she searches for the word and its neighbours using keyword-in-context, which shows five words either side of each match. She then hand-labels 40 matching sentences as framing skilling as an economic need, a social inclusion issue, or something else.

Next she tests a simple automated method against her labels. Where the two disagree, she reads the sentences and notices that the tool mistakes passages about unemployment for passages about skilling. That observation, and her explanation of why it happens, is the valuable part of the project. Her note ends with a limit (ten documents is a small sample) and a next step (add state-level documents). A reviewer reading this sees domain judgement, method and honesty about error.

Worksheet: describe your domain knowledge as an AI asset

Complete each line in writing. Aim for specific nouns and verbs rather than adjectives.

Turn the worksheet into a two-sentence pitch

Use this pattern: “I am a [programme] student who works with [type of text]. I can [specific judgement], and I have built [small project] to test how a model handles it.” Then add one line on what you plan to learn in the internship. Keep the whole pitch under 50 words for applications and profiles.

JNU humanities and social science students: placement-cell context for AI internship planning

JNU’s Placement Cell policy page says that institutions registered with JNUPC for student placement and internships are asked to upload their registration certificates. The page is an excerpt, and its year is not shown, so check the current version before relying on it. In practice, it suggests that the placement cell has a registration process for organisations offering internships through the university’s own channels.

What this does not tell you is how an online internship you find yourself, such as a self-applied AI programme, is treated by your school or centre. Nothing here confirms whether such internships need approval, whether they carry credit, how they are evaluated, or whether rules differ between schools and centres. Do not assume any of that either way. If you want the experience noted in your academic record, you need written confirmation first.

Also keep a separate record for your own use. Your project note, source sheet and certificate are useful in applications regardless of what the university decides to recognise.

Verification steps for JNU students before you commit to an AI internship

1. Open the JNUPC policy page and check its current text, date and whether it applies to your school or centre.
2. Email the placement cell and ask in writing whether students may do self-sourced online internships and whether any registration or intimation is required.
3. Ask your programme office or chairperson whether an internship can appear on your record, and request the answer by email.
4. Check your programme handbook or ordinance for any clause on internships, fieldwork or project work, and note the clause number.
5. Confirm your timing against your academic calendar so the 14 days do not clash with submissions, exams or fieldwork.
6. Save every reply, with date and sender, in one folder.

Building a portfolio without a CS degree: what to include and how to show it

• A one-page project note in plain language with a chart.
• A clean notebook or script with comments explaining each step in your own words.
• A source sheet listing every document, its origin and date.
• A short limits section that names what the project cannot claim.
• Your completion certificate, if you complete an internship, as supporting evidence rather than the centrepiece.

Avoid inflating the project. A modest, well-explained analysis is more convincing than an ambitious one you cannot defend in an interview.

Questions to prepare for when you interview without a technical degree

• “Why AI?” Tie your answer to a real text problem from your field, not to general excitement.
• “Walk me through your project.” Follow question, data, method, check, limit.
• “How did you know the output was right?” Describe your hand-labelled sample.
• “What did you get wrong at first?” Have one honest example ready, such as noisy PDF text.
• “What would you learn next?” Name one specific skill, such as evaluation metrics or working with larger corpora.

Frequently Asked Questions

Do I need to learn machine learning theory before starting an AI internship?

Not in depth. For a two-week beginner internship, comfort with Python basics, simple statistics and a clear idea of what a model’s output means is a more realistic target than theory. Learn concepts as your project demands them.

Which policy documents are suitable for a first text-analysis project?

Choose public documents from official websites that share a format, such as annual reports or scheme guidelines from the same body across several years. Consistent structure makes cleaning easier and comparisons fairer.

How many documents are enough for a 14-day project?

Eight to fifteen is manageable. The aim is to complete the full cycle of question, cleaning, method, evaluation and write-up, not to process a huge corpus.

Can I use ready-made tools instead of writing code from scratch?

Yes. Using existing libraries or tools is normal. What matters is that you can explain what the tool does, test it against your own labels and describe where it fails.

How do I handle documents that mix Hindi and English?

Treat this as a finding rather than a nuisance. Many tools perform unevenly on mixed-language text, so note where results look unreliable and say so in your limits section. If unsure, restrict your first project to one language.

Will my JNU school recognise an online internship I find myself?

That cannot be assumed. Rules on recognition, credit or evaluation were not verified here, so ask your programme office and the placement cell in writing and keep their replies.

How should I describe a humanities background on a CV for AI roles?

Lead with the text types you work with, the research methods you can transfer and your project, rather than listing courses. Use the worksheet’s stronger answer patterns as your wording guide.

Is a free internship worth doing if I am starting from zero technically?

It can be, if you go in with a defined project and a pacing plan. A free two-week internship with a certificate on successful completion gives structure and a deadline, but the value depends on what you build and can explain.

Related Artificial Intelligence guides

• Adamas University Students: Building a Recommendation System Mini Project During an Internship
• CMR University Students: Presenting an AI Internship Project to Non-Technical Managers
• IIMT University Students: Choosing a Dataset for an AI Internship and Checking Its Licence
• All Artificial Intelligence internship guides