From college (fresher), in 6 months
Final-year students and recent graduates with some coding, aiming for an FDE, Applied AI Engineer or AI Solutions Engineer role.
What already counts
- Data structures, algorithms and the coding practice you already did for placements
- Python or Java from coursework, and any SQL from a databases course
- College projects, internships or hackathons you finished, even small ones
- Explaining things to classmates, running a club or a fest: early proof you can work with people
What to add
- Code that runs unattended on real, messy data and other people's APIs
- One AI project on realistic data with an eval, not only a demo
- Any experience of discovery: talking to a real user and scoping something small
- Writing for non-engineers: a README, a one-page scope, an update
- Deploying and debugging something that real people use
1Python, data and APIs for real workWeeks 1 to 4
Get comfortable with the daily engineering of the job: scripts that read messy files, call APIs with paging and retries, and answer questions with SQL.
Show for it: A GitHub repo with a script that pages through a public API, cleans a messy CSV and writes a short report, plus a README showing sample input and output.
2Ship something and debug itWeeks 5 to 8
Learn to work in a shared codebase and run a small service where others can use it, then find problems through its logs.
Show for it: A small service running in a container on a free cloud tier with structured logs, and one merged pull request to an open-source project or a friend's real project.
3Applied AI with proofWeeks 9 to 14
Build an AI feature on realistic data and prove how good it is with an eval you designed, then make it safer.
Show for it: A search or question-answering project over a few hundred real public documents or products, with a 50-question golden set, a first score, one change and a better second score in the README.
4Discovery with a real userWeeks 15 to 18
Practise the customer half of the job: find a real person with a real problem, understand it, scope something small and write it down.
Learn
Running a discovery callFinding the real problem behind the requestDecomposing a vague problemScoping the first projectWriting a scoping documentStatus updates and written communicationPractise
Field Lab: Question list for a first call at Kaveri BankField Lab: Restate Dukaan's chatbot request as a problemField Lab: One-page scope for Raftaar's gate delaysCustomer Room: A chatbot for nursing college admissionsCustomer Room: "Use AI to reduce delays"Field Lab: Raftaar weekly update from raw notesShow for it: Notes from a discovery conversation with a shop owner, clinic or college office, a one-page scope they agreed to, and the first slice built and in their hands.
5Integration, privacy and rolloutWeeks 19 to 22
Connect to systems you do not own without losing data, keep personal data safe and switch things on in stages.
Show for it: The ticket triage take-home finished with masking, a review queue, an eval and a recorded five-minute walkthrough.
6Package your proof and get hiredWeeks 23 to 26
Turn your work into evidence a hiring manager can see in two minutes, and practise every round of the interview.
Learn
From fresher to FDEA portfolio that proves FDE skillsAn FDE CV and profileFinding FDE roles in India and abroadHow the FDE interview worksThe decomposition roundThe re-engineering and coding roundsBehavioural rounds and the project deep divePractise
Field Lab: Map a job posting to your skills and a six-week planCustomer Room: Triage for a hospital help deskMock interview: Design a company knowledge assistantField Lab: Rewrite five weak CV bullets for an FDE applicationField Lab: Write the first screen of a portfolio READMEField Lab: Ask an FDE you admire for a referralShow for it: A CV where every project line has a number, two portfolio projects with walkthrough videos, five practice rounds done, and 20 targeted applications with referral requests sent.