bonsai@stack:~$ man technologies/ai-ml
$ ai/ml
NAMEAI / ML
USAGE6 entries
PROJECTS8 projects
RELATED AREAAI & computer vision
## DESCRIPTION
We apply modern machine learning approaches to build intelligent systems, from research prototypes to production deployment on edge devices and in the cloud.
## TECHNOLOGIES_USED
[01]Edge→ On-device deployment without the cloud
[02]PyTorch→ And other modern frameworks
[03]R&D→ Research with measurable results
## HOW_WE_USE_THEM
ai/ml.capabilities.ts6 lines
- 01→ capability("Training classification, detection, and regression models");
- 02→ capability("Computer vision: YOLO, ViT, custom architectures");
- 03→ capability("NLP: language models, fine-tuning, multimodality");
- 04→ capability("Optimization: genetic algorithms, reinforcement learning");
- 05→ capability("Quantization and pruning for deployment on constrained hardware");
- 06→ capability("MLOps: reproducible experiments, model versioning");
## OUR_CASES.logtail -f cases →
[01]
→[02]
→[03]
→[04]
→[05]
→[06]
→[07]
→[08]
→

LogiFlex
# Logistics
+7%
goods turnover growth

Anomaly detection system
# Aviation / R&D
87%
detection accuracy

Yandex
# Tech / R&D
+8 pp
Precision@10

Eye-tracking system
# R&D / HCI
91%
accuracy on real-world data

Part inspection service
# Manufacturing
×7.5
faster analysis

Federated learning
# R&D
−18%
gradient bias

AI in education
# Education / R&D
3000+
learners

RAG knowledge base
# Finance / Banking
10+
document formats
## RELATED_EXPERTISE