There is no degree requirement. Students, recent graduates, self-taught engineers and career switchers may apply, but demonstrated ability remains essential. The employer requires practical Python API/JSON/scheduled scripting, Postgres SQL, structured debugging, written communication and work with imperfect documentation. Notebook-only Python experience is insufficient. A shipped Claude Code or Codex project is mandatory, including an explanation of incorrect agent output and how you detected it.
Hands-on Airtable, Clay, n8n, Instantly and Postgres experience is required. Preferred extras include Next.js/TypeScript, Windmill and other tools; candidates need not possess every preferred skill. Work is approximately three office days weekly, with more during onboarding, standard IST hours and some helpful US-time overlap. The employer explicitly says this is not a night shift.
The role involves a live migration from no-code workflows to code, Postgres and an internal application: validated LLM extraction, integrations, orchestration, outbound infrastructure, tooling and reliability. Interns work directly with the founder. Covered Claude Code and WeWork access are advertised. Extension or a longer-term role depends on performance; neither is guaranteed. Vacancy count, closing date, exact daily hours, payment schedule and later salary are undisclosed.
Read the official careers page before applying. The listing remained public on 5 November 2026, and its linked form displayed the matching internship and a submission control. This confirms an available application route, not an independently verified response time or a promise that any particular applicant will qualify.
The employer-linked Airtable application requires a full name, resume attachment, city, locality, email, LinkedIn profile, earliest start date and monthly salary entry. Required experience ratings cover Claude Code/Codex, n8n, an IDE, outbound email tools, Python, APIs/integrations and Airtable. Project/portfolio links and the written interest answer appeared without required markers. Nevertheless, the vacancy's project requirement still applies; an optional form field does not waive it.
The salary field uses INR formatting. Entering a requested figure does not establish an offer or replace the published range. Prepare accurate availability and location information, and ensure your resume describes your own work. Jobsii inspected the public form without entering applicant details, selecting answers, uploading files or submitting an application.
The advice below is editorial preparation, not an additional employer assessment. Start with a project you have actually shipped. Make a short account of the user problem, the component you owned, a failure you encountered and the evidence showing your correction worked. A modest functioning system with a clear explanation is easier to assess than a large architecture diagram whose components you cannot explain.
For the coding-agent discussion, choose a concrete mistake. Perhaps generated code treated a missing identifier as a valid record, or retried an operation that was unsafe to repeat. Show the relevant assumption, the observation that exposed it and your correction. Distinguish the agent's suggestion from your judgment. If the project was collaborative, identify your contribution instead of presenting teammates' work as entirely yours.
Use fictional records to rehearse a migration plan. Imagine moving a small workflow that receives company information, enriches it through an API and saves the result. First define the current behaviour and the fields that matter. Then describe how you would compare old and new outputs, reconcile missing records and decide whether a cutover is ready. This is a preparation scenario, not a disclosed employer task.
Consider a duplicate delivery. If the same input arrives twice, what prevents duplicate records or repeated charges? Explain an identifier, a database constraint and the behaviour you expect on retry. Consider a partial failure too: the provider returns successfully but the database write fails. A useful answer identifies what state persists and how a later attempt avoids repeating an irreversible action unnecessarily.
Use a mock provider or saved fictional responses for practice. You do not need access to a real client's data, an active outbound campaign or a paid enrichment account to explain your reasoning. Keep a short reconciliation table showing an input identifier, expected outcome, actual outcome and unresolved discrepancy. That makes the discussion concrete without pretending a simulated exercise was production experience.
Create a small set of invented inputs with deliberately awkward cases: a missing field, inconsistent formatting, an ambiguous company name and an unavailable provider. Define which outputs should be accepted, rejected or reviewed. Explain how you would retain enough context to investigate a failure without treating every generated answer as correct simply because it matches a JSON structure.
Separate completeness from accuracy. A pipeline can fill every field and still attach information to the wrong entity. Describe checks for identity, provenance and uncertainty, then explain how you would inspect a failed case. For cost reasoning, count attempted calls as well as successful records. A retry loop that eventually succeeds can still waste budget if its failure pattern goes unnoticed.
Clarify the office location, onboarding attendance, expected timezone overlap and weekly schedule before committing. If you are studying, compare the arrangement with compulsory classes and exams. Ask how the stipend within the advertised range is determined, when payments occur and whether expenses or deductions apply. Obtain written terms for any extension rather than assuming three months automatically leads to continued employment.
Discuss supervision and review: who checks changes, how production access is introduced and what an intern should escalate. Ask what good performance would look like during the initial weeks. These questions help you evaluate whether the actual responsibilities and support fit your current abilities. Recheck the official application form and vacancy before sending your materials, since hiring status and instructions can change.
Review the role details, deadline, and work mode before applying. Jobsii tracks the apply action so candidates can keep application history inside their dashboard.
