AI-augmented nearshore development agency for startup product builds
Katapult Labs is a Bogotá-based development agency founded in 2022 by venture-backed operators with 20+ years of startup experience. The tech stack reveals heavy investment in AI tooling (OpenAI, Anthropic, Claude, xAI) layered atop a modern full-stack foundation (React, TypeScript, Python, AWS), and active projects span AI-native surfaces—video generation, filmmaking automation, marketing workflows. Rapid hiring acceleration (46 roles posted in 30 days) skews heavily toward senior engineers and product talent, suggesting aggressive scaling of both delivery capacity and in-house product capabilities beyond pure agency services.
Notable leadership hires: Head of Design, Head of Marketing
Katapult Labs combines AI with nearshore software development to build products for startups and growing businesses. Based in Bogotá and hiring across Latin America (Brazil, Peru, El Salvador, Chile, Mexico, Uruguay), the company employs 11–50 people and uses a modern cloud-native stack centered on AWS, React, and Python. Beyond client delivery work, Katapult is building internal AI-native products and tools—including a content machine, video generation suite, and workflow automation platform—suggesting a dual operating model: agency services paired with productized AI offerings. Current pain points include stabilizing the platform, establishing design standards, and implementing systematic QA processes as the organization scales.
React, TypeScript, Python, Next.js, Node.js, AWS (Lambda, Fargate, Redshift, Glue), PostgreSQL, and Supabase for core infrastructure. AI integrations include OpenAI, Anthropic, Claude, and xAI. PostHog for analytics, Salesforce and HubSpot for CRM.
Client software projects alongside internal AI products: a content machine, AI video and script generation tools, filmmaking automation stack, marketing workflow automation, and rapid experimentation frameworks. Heavy focus on AI-first tooling and automation.
Other companies in the same industry, closest in size