
Deployment Engineer
Amazon-quality product recommendations for any e-commerce store.
Amazon has the best product recommendations. They use their historic sales data to create these recommendations. The problem though is that other e-commerce stores lack enough sales training data to do the same.
We've solved this problem. The Depict.ai algorithm increases overall e-commerce revenue by 4-6% for any store without needing any sales data at all.
Skills: JavaScript, Python
Do you want to join a fast-paced startup and work on deploying the bleeding edge of AI and frontend technology? We are looking for a Deployment Engineer with 3+ years of coding experience.
What you will be building: Self-contained JavaScript plugins that add Depict.ai's AI-functionality to e-commerce websites, scraping infrastructure, and data processing pipelines to ensure high-quality data is flowing into our Machine Learning algorithms.
We need someone who we can count on to:
Own: Deployments into new customers, including customized self-contained JavaScript plugins, scrapers and
configurations for our Machine Learning algorithms.
Teach: How to build tooling and processes that ensure the best possible maintainability, deployment times,
and latency.
Learn: How to hack into popular SPA-frameworks, architect integrations resilient to minor style changes,
optimize loading speeds, and configure Machine Learning algorithms to function in their best environment.
Improve: Efficiency and robustness of integrations operating inside evolving environments.
We expect you to know:
- JavaScript (TypeScript), CSS and HTML
- Python
Nice to have
- Tracking and analytics
- Node.js
- Cloud computing (AWS, GCP or Azure)
- Puppeteer
- Kubernetes
Within 1 month we expect you to:
- Understand the ins and outs of why we're approaching our integrations as we do.
- Have talked with existing clients to understand their pain points.
- Read the book "Third Party JavaScript", https://www.manning.com/books/third-party-javascript
- Get to know our existing Deployment Engineer's and how they're using our framework at the moment.
- Build out 3-5 integrations yourself to get a sense of how our existing frameworks work.
Within 3 months we expect you to:
- Iteratively make data pipelines faster and more efficient.
- Contribute to a functional low-code point-and-click interface that sets up tracking and analytics automatically for most sites (~80-90% coverage).
- Contribute to a functional low-code WYSIWYG-interface where one can reliably style and inject widgets into almost any site (~80-90% coverage).
- Make major improvements on top of existing framework that scrape high amounts of e-commerce data every day from multiple sites. Highly resilient against anti-bot and style changes.
- Contribute to framework for visual automated tests that can verify that our integrations are working as expected.
Within 6 months we expect you to:
- Develop framework to support a higher variety of widget types and functionality (search, category listing, pop ups, e.t.c).
- Increased efficiency of internal tools to sustain a low amount of deployment engineers while exponentially increasing the number of customers joining our platform.
- Refine software to make as much as possible towards self-serve.
Within 12 months we expect you to:
- Build out sufficient amount of tooling and support to be able to provide a no-code self-serve solution to integrate with Depict.ai, agnostic to which e-commerce platform it's using.
- Start exploring ways to increase our product offering and upsell customers to not only show product recommendations on e-commerce pages.
Our selection process is continuous and we may move forward with candidates on an ongoing basis. If you have any further questions about the role, please contact [email protected].
- Scrape e-commerce stores to fetch their product data (using Kubernetes (Argo), Puppeteer & Tor for IP-rotation)
- Fine-tune & iterate on the recommender system (hybrid model based on content fetched from specialized ConvNets and collaborative filtering)
- Tooling to auto-generate JS-snippets that can inject our product recommendations widget and collect analytics of how it performs (Chrome Extensions, Babel and styling with CSS)