Aquila — AI Data Production Platform
Aquila is a platform for running and evaluating AI data production work. It puts automated task assessment, sandboxed code execution, deployment automation, project management and workforce operations in one system, so teams can build, test and review work in the same place.
My Key Contributions
AI-Powered Task Evaluation
Built systems that use AI to review completed tasks, check requirements and give feedback before human review. Pluggable provider adapters drive live LLM calls with cost tracking and response comparison.
Secure Sandbox Execution
Runs submitted code in isolated AWS CodeBuild environments with a buildspec rendered per task, then tests the output and has AI assess the result. Reports are stored in S3.
Automated Deployment & Preview Generation
Spins up a temporary EC2 preview environment per task branch with docker compose, resolving AMIs through SSM and pulling secrets from Secrets Manager, then tears it down on a TTL. Reviewers can test finished work without setting anything up.
Inside the Platform
Secure AI Data Production Environment
A highly secure environment with an integrated coding IDE where contributors build datasets end-to-end (0 → 1) under strict, no-copy-paste workflows.
Workforce Management System
Personnel management, project management, engineering workflows and evaluation all live in one system rather than in separate tools.
Hiring & Training Pipeline
Every developer goes through a one-week pre-hire training phase, evaluated by the Key Engineering team, with hiring decisions based on measured performance.
Quality & Evaluation Infrastructure
Task creation, review, dataset validation and contributor evaluation run through the platform, so training data stays consistent.
Tech Stack
A FastAPI backend on clean architecture, deployed on AWS with automated sandboxing, AI evaluation and full tracing.
Core & API
Data
Cloud & Infra
AI
Auth & Security
Async & Jobs
Observability
Integrations
Are you interested?
Let's work together on your project.