MLOps2026
Runs durable, checkpointed SQL workflows inside PostgreSQL so long-running data and AI pipelines can resume after crashes without external orchestrators. Provides a SQL DSL, in-process background worker, and Postgres-backed state—useful for embeddings, ETL, scheduling, and fan-out jobs when you can install extensions.
AI Train2026
Provides a one-command CLI to fine-tune and post-train LLMs, with layer streaming that lets an 8B model be fine-tuned on a 4 GB laptop GPU. Auto-configures quantization, LoRA adapters, batching and evaluation gates, and supports export and serving workflows.
AI Model2026
Identifies and surgically removes the internal activation directions that trigger refusal behavior in large language models, with one-click options on a HuggingFace Space or a local Python API. Combines multiple extraction methods (SVD, whitened SVD, sparse autoencoders), reversible steering, and analysis-informed verification to quantify capability and refusal trade-offs.
AI Agent2026
Automatically evolves Hermes Agent skills, prompts, tool descriptions and code using DSPy + GEPA — mutating text via API calls, evaluating trace-based failures, and selecting variants that pass tests and human PR review. No GPU training required; runs cost roughly $2–$10 per optimization.
AI Agent2026
Lightweight, Markdown-only skill pack that lets LLM agents autonomously run ML research workflows—literature survey, idea discovery, cross-model review loops, experiment automation and paper writing—designed for Claude Code, Codex CLI, Cursor and local model setups.
AI Agent2026
Gives the pi terminal AI agent an autonomous experiment loop: propose code changes, run benchmarks, record metrics, auto-commit improvements and revert regressions. Ships a live widget/dashboard, MAD-based confidence scoring, hooks and backpressure checks — made for iterating on speed, bundle size, training loss and build times inside a terminal workflow.
AI Agent2026
Turns a single research idea into runnable experiments and a conference-ready paper by orchestrating an LLM-driven end-to-end workflow (literature → design → code → sandboxed runs → analysis → writing). Provides human-in-the-loop checkpoints, domain-specialist executors, and multi-layer citation verification.
AI Agent2026
Runs an autonomous self-improvement loop where a meta agent crafts a task-specific agent, a target agent executes trials, and a feedback agent updates both harness (code) and model weights—provider-agnostic profiles with reproducible runs and a live dashboard.
AI Agent2026
Provides a CLI and skill suite that lets coding assistants scaffold, evaluate, and deploy ADK-based AI agents on Google Cloud. Integrates eval pipelines (generate/grade), deployment infra and CI/CD scaffolds, observability, and Gemini Enterprise publishing workflows.
AI Agent2026
Monitors and detects risky behavior in enterprise AI agents via high-fidelity telemetry, security benchmarking, and a two-tier detector. Comprises ADR Sensor, ADR-Bench, and ADR Detector; deployed in production at Uber and validated on public benchmarks.
AI Infra2026
Provides an end-to-end platform to evaluate, observe, protect, and optimize LLM and AI agent deployments. Integrates OpenTelemetry tracing, 50+ evaluation metrics, agent simulations, an OpenAI‑compatible gateway, and guardrails; self‑hostable under Apache 2.0.
AI Infra2026
Defines OpenTelemetry semantic conventions for generative AI telemetry — spans, metrics, and events for GenAI clients, the Model Context Protocol (MCP), and provider-specific integrations. Includes YAML models, human-readable docs, and reference implementations to standardize observability across GenAI deployments.