Dixamix
Selected work

Systems we've shipped to production.

Each engagement reflects deep ownership across architecture, engineering and deployment — designed for the long run, not for a demo.

Computer Vision for Custom Orthopedic Insoles
Case · 01

Computer Vision for Custom Orthopedic Insoles

Challenge

Replace specialized in-store scanners with a smartphone-only flow that captures clinically usable foot biometrics for custom orthotic insoles.

Solution

A computer vision pipeline using YOLO models for keypoint and structure detection, combined with biometric extraction logic and a mobile capture experience.

Outcome

A scalable, accessible measurement platform — designed for production reliability and clinical fidelity.

YOLOPyTorchFastAPIMobile captureOpenCV
Healthcare Platform with Medical RAG Assistant
Case · 02

Healthcare Platform with Medical RAG Assistant

Challenge

Deliver a healthcare platform with conversational support, scheduling logic and trustworthy clinical knowledge retrieval.

Solution

A medical RAG assistant with vector search, intent routing and guardrails — integrated into a Next.js platform with scheduling and patient flows.

Outcome

A safety-aware AI assistant embedded in a real clinical product, designed for maintainability and audit.

Next.jsVector DBRAGIntent routingHealthcare
Business Systems on AWS & Kubernetes
Case · 03

Business Systems on AWS & Kubernetes

Challenge

Unify Node.js services and an Odoo ERP into a single, container-orchestrated platform with operational discipline.

Solution

Containerized services on Kubernetes (AWS), GitOps deployments, observability and a clean separation between business logic and integrations.

Outcome

A reliable internal platform that operations teams can trust and engineering teams can extend.

Node.jsOdooDockerKubernetesAWS
MLOps & Data Lakehouse Platform
Case · 04

MLOps & Data Lakehouse Platform

Challenge

Stand up production data and ML infrastructure across streaming, batch and feature-serving — without accumulating tech debt.

Solution

An EKS-based MLOps platform with feature store, Kafka/Iceberg lakehouse, dbt and Airflow stacks, MLflow registry and end-to-end observability.

Outcome

A coherent foundation enabling data, analytics and ML teams to ship with confidence.

EKSKafkaIcebergAirflowdbtMLflowFeast

Working on something similar?

Let's scope it together.

Start a project →