We build and improve optimal technology stacks according to each situation, enabling us to develop and scale products at maximum speed.
Frontend
We build rich web frontends for chat administration and dashboards using JavaScript/TypeScript and React.
Backend
Our services use primarily Go, with Python and Elixir for specific features. We follow microservices architecture across domains like conversation flow, user data, and analytics, communicating via gRPC and message queues. We use cloud task queues for bulk operations and monitor with Prometheus, Grafana, and OpenTelemetry.
AI / ML
We use PyTorch and TensorFlow for training models in intent detection, language understanding, and content recommendation. Our stack integrates LangChain and Haystack for agent orchestration and semantic search, enabling rapid prototyping and deployment of tailored AI agents.
Data / Analytics
We use Google Cloud Platform with BigQuery as our data lakehouse, storing event and conversation data in Apache Iceberg format from multiple sources. Our ELT pipelines leverage Airflow, PySpark, Scala, and dbt for data transformation and contracts, with reporting through GoodData and Tableau.
Infrastructure
We deploy on GCP using Docker containers on Kubernetes with ArgoCD for deployment management. Infrastructure is defined in Terraform, with CI/CD pipelines automated through GitHub Actions.
AI coding / Dev Tools
To continuously adapt to rapidly changing technological environments, we flexibly adopt the best tools and prioritize those that contribute to developer productivity improvement.
We embody both technical excellence and an agile development culture.
Quality is everyone’s responsibility. We do peer code reviews on all pull requests to maintain standards and share knowledge. We have an automated test suite (unit, integration tests) that runs on each CI pipeline – catching issues early. Our QA engineers build automated tests for chat flows to ensure new releases don’t break existing conversations.
“You build it, you run it.” Developers at ZEALS take ownership of their services beyond just writing code. This means setting up proper monitoring and being on-call in rotations for the services you know best.
We’re always refining our processes. In retros, team members openly discuss what can be better – whether it’s our sprint structure or the onboarding process for new hires.
ZEALS encourages innovation and experimentation. We periodically organize hackathons where engineers can team up to prototype crazy ideas. We also allow engineers to spend a portion of time (if workload permits) on learning or trying out new tech that could benefit the product.
We broaden our expertise within the team through Tech Talks and pair programming sessions. For instance, team members who develop deep knowledge of new GCP services or testing approaches share their insights through presentations, working to raise the collective competency of our entire engineering team.
