Senior Backend Engineer
Kembangkan server MCP untuk menyediakan data merek terstruktur ke AI/LLM
Bangun infrastruktur backend untuk Brandfetch, platform aset merek. Fokus pada skalabilitas dan kecepatan pencarian vektor logo.
Kenapa Menarik?
Dapatkan ekuitas di perusahaan yang menguntungkan dan berkembang pesat
Skills Wajib
Konteks Indonesia
- Overlap Jam Kerja:
- Sedikit overlap dengan jam kerja Jakarta
Keywords
Lihat Deskripsi Asli dari HN Who is Hiring
Deskripsi asli dari HN Who is Hiring
Brandfetch ( https://brandfetch.com ) | Senior Backend Engineer | Remote | Full-time Brandfetch is the internet’s canonical source for logos and brand assets. Think GitHub for brand identity: a brand repo that auto-indexes public assets, lets brand owners “commit” changes once, then propagates them—instantly and immutably—through CDNs, plugins, and APIs used by companies like Google, Canva, Salesforce, and soon … everywhere. Problems you'd work on: - Nightly curation bot auto-fixing outdated assets across 52M+ repos - Next-gen ingestion and partition-key schema to leap from millions to billions of assets - MCP server that exposes structured brand data to any AI agent or LLM, letting them fetch brand context on demand - Vector search: surface any logo with ≥90% similarity in <50ms - Real-time metering that pings Stripe the instant a user tops their quota Stack: Node.js, TypeScript, Python, AWS (Lambda, DynamoDB, CloudFront, S3, CloudFormation), Serverless. You : - 8+ years of experience in backend software development. - Proficiency with AWS tools (Cloudformation, API Gateway, DynamoDB, Lambda, Cloudfront, S3, IAM, OpenSearch). - Experience with building REST APIs. - Ability to write well-documented, clean code. - Experience with building DevOps pipelines using Serverless. - Great communication skills (we work across multiple timezones and document everything). - Experience with Machine Learning. - Timezone: UTC+0 to UTC+9 Competitive salary + equity in a profitable, growing company. apply: career at brandfetch.io (no recruiter please).
