Multifunctional AI Agents for SaaS
Publicada el 2026-08-02
Descripción de la oferta
I’m building a green-field SaaS startup and need a set of AI agents that will sit at the heart of the platform. Each agent must seamlessly blend Natural Language Processing, Machine Learning and Predictive Analytics so it can absorb both structured and unstructured data, learn from every interaction and surface actionable, forward-looking insights for our users in real time. What I need you to do • Design the overall agent architecture as modular micro-services that we can scale independently. • Implement and train the first production-ready models on the sample data I provide, then document how my team can retrain them as fresh data arrives. • Expose the agents through a clean REST or GraphQL API and, where it makes sense, a conversational chat interface. • Package everything with unit tests and deployment scripts (Docker/Kubernetes preferred) so we can drop the stack straight into our CI/CD pipeline. Acceptance criteria 1. NLP layer must reach at least 90 % intent-classification accuracy on the supplied test set. 2. ML component retrains without manual intervention and version-controls its models. 3. Predictive module produces forecasts whose MAE sits within an agreed tolerance we’ll define together. 4. Code is clean, comprehensively commented and delivered in a private Git repository. In your proposal, briefly outline the libraries or frameworks you’d lean on—Hugging Face Transformers, scikit-learn, TensorFlow, Prophet, or anything else you feel is fit—and your estimated timeline for an MVP. I’m ready to start this week, so feel free to fire over any technical questions that will help you scope the build accurately.
Skills
- REST APIs
- NLP
- Microservicios
- Machine Learning
- AWS
- scikit-learn
- Java
- Python
- Kubernetes
- TensorFlow
- Docker
- Hugging Face
- GraphQL
- CI/CD
- Git
Fuente original: freelancer