Senior Engineer · AI Resume & Coaching

Bewerbung.AI

Building Germany’s AI-powered application platform — résumé, cover letter, coach, and Bewerbung flow.

Timeline

2023 – 2024 (consulting)

Location

Remote · Berlin, Germany

Stack

4 areas · 20 tools

Scope

6 areas · 7 highlights

Overview

Bewerbung.AI is a German-market job-application platform that handles the full Bewerbung process: AI résumé (Lebenslauf), agent-driven cover letters (Anschreiben), job-fit checks against postings, and structured coaching through interview prep and follow-up.

I owned slices across editor UX, backend APIs, agent flows, PDF export, analytics instrumentation, and cost-aware AI metering — the product is document-heavy where export fidelity and editor responsiveness are the promise.

My role

Built the React + RSPack editor (V2 resume/cover-letter architecture with Redux Toolkit and RTK Query) and Express/Mongo backend powering the studio.
Shipped agent-side Anschreiben generation — dedicated cover-letter flows and hooks so agents personalise German tone, structure, and role-specific arguments from resume context.
Designed job-fit and coaching agents: discovery, gap analysis, role-match grading against postings, and interview prep with concrete edit suggestions.
Moved heavy PDF export to AWS Lambda jobs and hardened template rendering so links, rich text, and layout survive editor → API → export round-trips.
Set up product analytics (Statsig, Mixpanel, Sentry) with deferred loading and Rspack vendor chunks so instrumentation does not dominate cold start.
Implemented proactive improvement emails, Stripe credits for AI usage, and eval-driven prompts to protect quality and unit economics.

Stack

Frontend

ReactRSPackTypeScriptTailwindRedux ToolkitRTK Query

Backend

Node.jsExpressMongoDBPM2

AI

Cover-letter agentsJob-fit & coaching agentsOpenAI / AnthropicEval & prompt versioning

Infra & product

AWS Lambda (PDF)Stripe creditsPostmarkStatsigMixpanelSentry

Highlights

Shipped a full Lebenslauf editor with AI authoring, live preview, section-level rewriting, and multi-template PDF output.
Built agent-side Anschreiben (cover letter) generation with V2 cover-letter state parity — German formality and role-specific drafting.
Delivered job-fit check: coaching agents grade applications against role requirements with gap analysis before candidates apply.
Hardened export fidelity — publication links, structured fields, and layout preservation across desktop/mobile and many templates.
AWS Lambda PDF jobs isolate heavy renders from the Express API; Rspack bundle tuning and deferred analytics reduce main-thread and infra cost.
Analytics setup: Statsig experimentation, Mixpanel events, Sentry — loaded on idle paths without blocking editor cold start.
Stripe paywall, credits metering, and proactive Postmark nudges that score applications and ship actionable improvements.

Outcomes

Made high-quality German Bewerbungen accessible to candidates who don’t have access to expensive coaches.
Reduced time-to-application from hours to minutes with agent-assisted Anschreiben and trustworthy exports.
Raised application quality through job-fit checks, coaching agents, and proactive improvement loops.
Protected SaaS unit economics with credit-gated AI, eval-driven prompts, and cost-aware frontend/analytics loading.

FAQ

What is Bewerbung.AI?

Bewerbung.AI generates AI Lebenslauf and Anschreiben, runs job-fit checks against postings, and coaches candidates through German-market conventions — with proactive emails when applications can improve.

What did you build on the agent side?

Cover-letter agents for Anschreiben drafting, coaching agents for job-fit and gap analysis, and a proactive email agent that surfaces highest-impact edits — all behind Stripe credits and eval-driven prompts.

Why MongoDB and Express on the backend?

Application data is naturally document-shaped (résumés, drafts, coaching transcripts) and Express keeps the surface small. Heavy PDF work runs on Lambda so the API stays responsive.