mindcurvature --mission
We help companies build AI systems that last — from pragmatic prototypes to production-grade platforms.
Rooted in fundamentals. Driven by execution
mindcurvature --services
🤖
End-to-End AI Solutions
ML/AI inference, RAG systems, AI agents, GPT-like chats.
🚀
Rapid MVP Delivery & Scale
We architect and deliver data-intensive, high-performance systems that grow with you.
💰
Cost & Performance Optimization
Launch fast with lean, focused prototypes that solve real problems.
⚡
High-Performance Applications
Optimize for performance, scalability, and cost without sacrificing reliability.
🔄
Production-Grade MLOps/LLMOps/DataOps
Bring structure, repeatability, and automation to your AI lifecycle.
mindcurvature --team
Anton Plotnikov
ex-Shopify / ex-Accenture
15+ years of hands-on experience
AI Carpenter — builds solid, scalable, elegant foundations for ML and software systems
- Software Engineering
- MLOps Engineering
Aleksandr Beloglazov
ex-Apple / ex-Zalando
12+ years of hands-on experience
Data Plumber, ML Welder — gets the messy data flowing and ML pipelines humming
- MLOps Engineering
- DevOps Engineering
mindcurvature --ml-engineering
- Model Types: ASR, TTS, LLMs, embeddings, image & video models
- Inference & Serving: NVIDIA Triton, TensorRT, ONNX Runtime, vLLM, SGLang, DeepSpeed, TorchServe, KServe, BentoML, Ray Serve, OpenLLM, Hugging Face TGI
- Experimentation: MLflow, ClearML, Weights & Biases, neptune.ai
- Orchestration & Scaling: Ray, KubeRay, Slurm, Kubeflow Pipelines, Flyte, Metaflow, Airflow
- GPU Infrastructure & Optimization: CUDA, cuDNN, TensorRT
- Versioning & Registry: MLflow, DVC
- Languages: Python, Golang, Java, Scala, JavaScript, TypeScript, C++
mindcurvature --devops
- Cloud Platforms: AWS, GCP, Azure, on-premise, hybrid
- Kubernetes: Helm, Kustomize, Argo CD, Istio
- Infrastructure as Code: Terraform, Pulumi, Ansible
- CI/CD: GitHub Actions, GitLab CI, Argo Workflows
mindcurvature --observability
- Metrics: Prometheus, VictoriaMetrics
- Logs: ELK (Elasticsearch, Logstash, Kibana), Loki, FluentBit
- Tracing & Profiling: OpenTelemetry, Jaeger, Tempo, Pyroscope
- Dashboards: Grafana, Superset, Metabase
mindcurvature --data-engineering
- Batch/Stream Data Processing at Scale: Apache Flink, Apache Spark, Apache Pulsar, Apache Kafka
- Analytical Datastores/Platforms: ClickHouse, Snowflake, Databricks
- DataOps & Workflow Orchestration: Apache Airflow, Prefect
- ETL: dbt
mindcurvature --contacts
Email:
contact@mindcurvature.com
Location:
Suite 17, Beacon South Quarter
Sandyford Business Park
Dublin, Ireland