Anton Ganichev
Computer Vision Engineer & AI Consultant
I do research and prototyping in Computer Vision and generative models (images/video).
I help research and engineering teams apply AI in practice,
combining hands-on implementation with mentoring and knowledge transfer.
My recent focus is conditional generation and diffusion-based image/video editing.
26+ years in IT / software development.
What I deliver
- Hands-on prototypes/baselines (notebooks/scripts) for CV and generative workflows, built for reproducible iteration
- Method and architecture reviews: selecting viable approaches, clarifying assumptions/risks, defining an experiment roadmap
- Mentoring and technical guidance: deep learning for scientific research (PhD-level) + project guidance for interns/juniors (code + experiment reviews)
Primary focus
- Controllable generation. & editing (diffusion-based): spatial/reference conditioning; img2img/inpainting/inversion workflows
- Video CV: detection + tracking pipelines, evaluation pitfalls, and practical integration patterns
- Segmentation & multi-task learning: modern architectures, losses, mask processing, dataset/annotation nuances
Background (familiar topics)
- Deployment basics (Docker; general awareness of ONNX/TensorRT)
- LoRA / parameter-efficient tuning (conceptual + practical exposure)
- Depth estimation, VLM-based anomaly detection, data unlearning (study-level)
Primary tools: Python · PyTorch · Linux · Git · Docker · OpenCV · torchvision · Transformers · Diffusers · Hydra
Used as needed: Ultralytics YOLO · Stable Diffusion / Flux · ControlNet · ComfyUI · CLIP
Work format
Remote · contract / consulting · training / curriculum development.
English: B2 (Upper-Intermediate), spoken.
Contact: gan4x4@gmail.com.