Gemma Tuner Multimodal vs Seeknal

Gemma Tuner Multimodal is an open-source fine-tuning toolkit for Google's Gemma 4 and Gemma 3n models that runs entirely on Apple Silicon using PyTorch with Metal Performance Shaders (MPS) backend — no NVIDIA GPU or cloud infrastructure required. It supports LoRA training on multimodal inputs: audio, images, and text simultaneously, using local CSV files or streamed from Google Cloud Storage or BigQuery. The tool targets the growing segment of developers who own M-series Macs but have been locked out of fine-tuning workflows that assume CUDA availability. Gemma 4's architecture is particularly well-suited to this use case: its 4B multimodal variant (designed for on-device deployment) trains efficiently on M3 Max and M4 Pro hardware within the available unified memory constraints. Primary use cases include medical transcription fine-tuning (audio → text with clinical terminology), visual QA systems (image + text → structured response), and private on-device pipelines where cloud API calls are prohibited by compliance requirements. The project fills a specific niche that Google's own fine-tuning documentation doesn't cover well for Apple hardware.

Seeknal is a Data & ML CLI designed for teams running agent-driven data pipelines. The core workflow follows three verbs: Organize (define pipelines in YAML or Python), Expose (materialize data to PostgreSQL and Apache Iceberg), and Action (query and transform data in natural language). It uses a draft, dry-run, apply progression that gives teams control before changes hit production. The natural language query layer is what sets Seeknal apart from standard data pipeline tools. Instead of writing SQL to explore a freshly materialized table, you describe what you want — and Seeknal translates that to the appropriate query against your Postgres or Iceberg target. The combination of structured pipeline definition (YAML/Python) with flexible natural language exploration is designed for the reality that data teams include both engineers who want explicit control and analysts who want fast iteration. The 'built for the agent world' framing reflects a genuine architectural choice: Seeknal's API is designed to be called programmatically by AI agents, not just by humans with keyboards. This matters because data pipeline management is increasingly something agents need to do autonomously — fetching fresh context, materializing results, and querying outputs — without human intervention at each step. Seeknal launched on Product Hunt today targeting teams that have adopted agentic workflows but still treat their data infrastructure as human-operated.