Vertex AI and Gemini pipelines
Structured transcription and generation with the model picked per job by length and difficulty to control cost.
Google Cloud development
Google Cloud and Vertex AI for production AI pipelines.
We use Google Cloud mainly for AI. The lecture transcription pipeline in our case studies runs Gemini Pro, Flash and Flash-Lite through Vertex AI, stores media in Google Cloud Storage and reaches 99% precision on Hebrew and Aramaic, with model selection cutting AI cost by up to 60%.
Google Cloud at OSKIFrom case studies
Client projectsPublished case studies that used Google Cloud
2projects
Where in the stackAI, Storage
2layers
IndustriesMedia and EdTech, Cross-industry
2industries
Most recent projectAI wired into CRM, ERP and support to cut manual work across operations
2026year
What we build with Google Cloud
Where Google Cloud earns its place.
The kinds of work we use Google Cloud for, based on the projects below rather than a list of everything the technology can do.
Storage and media processing
Google Cloud Storage alongside S3 for multi-cloud media pipelines.
AI services for business
Google Cloud AI next to other ML tools in automation that plugs into CRM and ERP systems.
2 client projects
How Google Cloud fit into each project.
Which part of the system it ran, and what it worked with. Every row links to the full case study.
Transcribing Hebrew and Aramaic lectures at 99% precision with Gemini
AI · StorageGoogle Vertex AI, Gemini Pro, Gemini Flash, Gemini Flash-Lite, Google Cloud Storage
Together withPython, FastAPI, FFmpeg, FFprobe, AWS S3, Kubernetes
AI wired into CRM, ERP and support to cut manual work across operations
AIGoogle Cloud AI
Together withFastAPI, Flask, MongoDB, PostgreSQL, TensorFlow, PyTorch
Google Cloud case studies
Results from these projects.
Each one with the client’s problem, our approach and what changed.
Media and EdTech
AI2025
Media and EdTech
Transcribing Hebrew and Aramaic lectures at 99% precision with Gemini
A Kubernetes pipeline that turns hours of mixed-language lectures into timed transcripts, subtitles, summaries and HLS video.
99%precision on Hebrew and Aramaic
300+concurrent transcription jobs
AI2025–2026
Cross-industry
AI wired into CRM, ERP and support to cut manual work across operations
AI services connected to existing CRM, ERP and marketing tools: 24/7 chatbots, sentiment analysis, fraud alerts and predictive insights.
2xbusiness efficiency
80%data processing accuracy
Related services
Google Cloud is a tool. These are the jobs it does.
We do not sell hours of a technology. We take on a result, and choose the stack with you. These are the services where Google Cloud usually comes in.
Technologies
Other technologies we have shipped with.
Each one backed by at least one published client project.
Backend
Cloud & DevOps
Google Cloud FAQ
Google Cloud questions we hear.
Anything missing? Send us your question and an engineer will answer, usually within one business day.
Contact usWhy Gemini for transcription?+
For long, mixed-language audio it gave the best balance of precision and price in our tests. We still pick models per job, and the pipeline can switch.
Can you combine Google Cloud with AWS?+
Yes. The transcription pipeline uses both Google Cloud Storage and AWS S3, with Kubernetes running the workers.
How do you keep AI costs under control?+
By choosing the model per task, caching results and splitting long inputs sensibly. Model selection cut AI and API cost by up to 60% on the transcription project.
Start a project
Tell us what you are building with Google Cloud. We come back with a plan and a price.
A spec, a Figma file or three paragraphs is enough to start.
- First reply · from an engineer, not a sales rep
- 30-min call · with Roman, our CTO, to clarify scope
- After the call · a written proposal with plan, team and price
Roman Oleksuk, CTOPrefer to talk? Pick a slot in Roman's calendar