Machine Learning Researcher

Extend is building a modern document processing cloud. We're on a mission to transform how the world works with unstructured data. As a Machine Learning Engineer at Extend, you’ll be responsible for building state-of-the-art document processing infrastructure. The work you do will impact every single customer across the millions of documents that our system ingests and processes every month. We’re looking to hire engineers who think creatively about problems, keep up to date on the latest in AI research, push for real-world customer impact (not only theory), and are excited about helping businesses tackle one of their most painful problems with AI. WHY YOU SHOULD CONSIDER JOINING * We've grown revenue 400% last year to several million ARR, with our growth continuing to accelerate * We have a product loved by users and being used in mission-critical flows at enterprises (Square, Zillow, Chime) and startups (Brex, Mercury, Checkr), and many more * The market for document processing has expanded 1000x due to LLMs, and all existing solutions are low NPS * We're punching well above our weight, supporting customer and revenue metrics with half the team size of other startups — everyone joining at this stage will have outsized impact * We've raised our series A & de-risked the business, but there's tremendous upside ahead; now's the perfect time to join. * You'll be joining a talent dense team (e.g. former founders, world record holders) operating in a high performance culture, in-person in NYC, with high equity ownership * We're tackling novel, unsolved problems at the intersection of LLMs, complex data processing, and human ↔ AI interfaces * You'll get a front row seat to how a successful startup is built with plenty of leadership opportunities as we grow quickly We're a lean team, and our customer base is growing too fast for us to keep up. We're looking for talented engineers to join our team and execute on an ambitious roadmap. ROLE * Training and deploying SOTA vision models for document processing * Design novel LLM techniques for increasing the complexity of use cases and data streams that Extend can be applied to * Monitor existing models in production, understand what’s working well (and what isn’t), and run experiments to solve those issues * Have complete ownership over the work you do — as a founding team member, you’ll have the opportunity to truly own large areas of product and engineering and have direct relationship with customers. * Work directly with the CEO, CTO, and other founding members and help build out our team. BENEFITS Beyond the below, we care deeply about each of our candidates on a personal level. We take the time to understand your needs to craft a package designed for you: * Insurance - 90% health insurance premium coverage * Relocation - for candidates based outside of NYC * Unlimited PTO policy - we trust everyone like owners, take what you need * Daily lunch & snacks covered (plus dinner if you're in the office late) - you'll never go hungry and you help pick the snacks * Unlimited token / tooling access - no handicaps on your productivity * Learning and development investment - we invest heavily in our team and support your growth Extend is an Equal Opportunity Employer. We consider all qualified applicants for employment without regard to race, color, religion, sex, sexual orientation, gender identity or expression, national origin, age, disability, genetic information, citizenship status, marital status, pregnancy, protected veteran status, or any other characteristic protected by applicable federal, state, or local law.

WARUM DU DICH FÜR UNS ENTSCHEIDEN SOLLTEST? Du arbeitest in einem interdisziplinären Team von Mitarbeitern mit einem soliden Hintergrund in Mathematik und Statistik. Wir bieten flexible Arbeitszeiten (Voll- und Teilzeit) und haben ein schönes Büro mit gutem Kaffee in Berlin Schöneberg. Wir bevorzugen ein hybrides Arbeitsmodell, sind aber auch offen für mobiles Arbeiten. Wir glauben an die Wissenschaft und unterstützen Sie bei der Veröffentlichung Ihrer Forschungsergebnisse. Hier sind kurze Beschreibungen von einigen Projekten, an denen wir gearbeitet haben. Schätzung der Anzahl von Sonnenkollektoren, die auf ein Dach passen (Computer Vision): Anhand eines Satellitenbildes und eines Bodenbildes eines Hauses sollen bestimmte Elemente eines Daches (einschließlich Hindernisse, Dachgauben usw.) automatisch erkannt werden, um herauszufinden, wie viele Sonnenkollektoren darauf passen. Dies beinhaltet die Ableitung von 3D-Informationen aus 2D-Bildern, um die Dachneigung zu ermitteln. Erkennen, Klassifizieren und Vorschlagen der rechtlichen Wirksamkeit von Textabschnitten (NLP): Automatisches Durchgehen von Tausenden von Rechtsdokumenten mit dem Ziel, bestimmte Absätze zu klassifizieren und ihre rechtliche Wirksamkeit zu überprüfen. Dies beinhaltet die Konvertierung von Scans in Text, die Entwicklung eines Etikettierungsschemas (Problemmodellierung) und die automatische Erkennung verschiedener Absätze, bevor die Inferenzaufgabe in Angriff genommen wird.