AI Engineer - Intern

COMPANY OVERVIEW ClickPost is a post-purchase and shipping intelligence platform for enterprise and mid-market e-commerce brands. We own the part of the order lifecycle that decides whether a customer buys again: tracking, delivery experience, returns, and the data that turns WISMO tickets and RTOs into revenue retention and upsell. We work with e-commerce brands across India, the US, and EU, and we're mid- expansion into North America right now, which is where this role sits. WHY THIS INTERNSHIP IS UNIQUE This isn't a rotation program. You report to one person: the Founding Member running Global Expansion for NA and EU, who owns a real pipeline and a $300K new-ARR number this year. No shadowing, no busywork, no toy dataset. What you build gets used in live by the ClickPost team & the customers. WHAT TO EXPECT WHEN YOU JOIN • Direct daily access to a founding-level GTM operator, not a manager three layers removed. • Fast decisions. No committee sign-off to ship something. • Ownership of a real system from week one, not tickets. • Weekly reviews tied to pipeline and reply-rate impact, not activity or hours logged. • High ambiguity, high autonomy. You'll be handed a business problem, not a spec. RESPONSIBILITIES • Build AI agents that scan D2C and e-commerce brand signals (funding, hiring, tech-stack changes) to flag active buying intent. • Turn those signals into personalized SDR outbound at the account level, not templated blasts. • Automate the manual GTM work currently done by hand: research, enrichment, sequencing, reporting. • Build internal copilots and research agents for Sales, Marketing, and Market Research. • Ship fast, measure against pipeline and reply-rate metrics, iterate weekly. EXAMPLE PROJECTS YOU'LL OWN • Signal-to-Sequence agent: ingests funding, hiring, and stack signals for named accounts and outputs a ranked list with a drafted first touch. • SDR research copilot: given a company URL, returns an account brief (pain points, detected 3PL/carrier stack, relevant proof points) in under 60 seconds. • Outbound QA agent: scores draft outreach against our rules (no generic opener, quantified value in the buyer's numbers, one clear next step). • Competitive signal tracker: monitors competitor moves publicly and pushes a summarized digest. • Internal RAG copilot: over collateral, case studies, and pricing, so SDRs self-serve answers instead of pinging the team. REQUIRED TECHNICAL SKILLS • Strong Python. You should be comfortable writing production-adjacent code, not just notebooks. • Working knowledge of LLM APIs (OpenAI, Anthropic, or similar). • Prompt engineering fundamentals: few-shot design, structured output, basic evals. • Bonus if you have experience building or orchestrating AI agents (LangChain, LlamaIndex, or your own framework). • MCP (Model Context Protocol) fundamentals, or the ability to pick it up fast. • REST API integration: auth flows, webhooks, rate limits. PREFERRED QUALIFICATIONS • A hackathon project, personal build, or open-source contribution involving LLMs or agents. Show us, don't just tell us. • Exposure to workflow automation tools (n8n, Zapier, Make). • Exposure to sales or GTM tooling (CRM, sequencing platforms) is a plus, not a requirement. • Currently pursuing or recently completed a degree in CS or Engineering. Demonstrated build ability matters more than the degree. WHAT MAKES SOMEONE EXCEPTIONAL FOR THIS ROLE • Builds before asking permission. • Treats it doesn't work yet as a starting point, not a stop sign. • Cares whether the output actually gets used by the team, not just whether the code runs. • Comfortable being handed a business problem instead of a spec. • Ships an ugly v1 in days over a polished v1 in weeks. LEARNING & MENTORSHIP • Direct 1:1 mentorship from a Founding Member & senior leaders. • Weekly build reviews: what shipped, what moved the metric, what's next. • Real exposure to enterprise GTM: deal math, ROI modelling, buyer psychology, live deal reviews. • No layers between you and the person who owns the business outcome. SUCCESS METRICS DURING THE INTERNSHIP • Qualified signals surfaced that convert into SDR-worked accounts. • Reply rate and meeting rate on AI-assisted outbound versus baseline. • Hours saved per week on manual research and enrichment across the team. • At least one system running in production, not a prototype, by the end of the internship. COMPENSATION & PERKS • Stipend: 35-50k per month. • Duration: 6 months, starting immediately. • Full-time conversion possible based on performance on or before 6 months internship duration closure. • A direct letter of recommendation, plus a real portfolio of shipped, revenue-linked AI systems, not a certificate of participation. APPLICATION PROCESS • Apply with your resume and one thing you've built with AI or LLMs. Send a link, not a description. • Practical AI assignment based on a real ClickPost problem • Technical discussion, Final discussion & then Offer (No competitive programming rounds.) • We care more about how you think, build, and how you solve problems.