Product roadmap

What exists now, what comes next, and what’s out of scope

Current capabilities

·Six-stage AI lifecycle with shared contracts
·Strategy intake, scoring & build-path recommendation
·Data readiness handoff
·Build/RAG lab — retrieval modes (BM25 / vector / hybrid / re-rank)
·Agent & tool-calling mechanics (schemas, boundaries, approvals, misuse evals)
·Training / fine-tuning readiness — decision memo, dataset readiness, overfitting & generalization
·Operate / AI Ops / MLOps spine
·Govern live evidence loop & decision engine
·Realize risk-adjusted ROI engine
·Model-internals explainer (transformers, attention, embeddings, framework placement)

Next technical upgrades

·Real vector retrieval / vector database integration
·Persistent eval run history
·Deeper telemetry integration
·Real tool integrations (APIs, workflow/ticketing engines)
·Real labeling tool + model registry + training pipeline integration
·FinOps: cost chargeback & unit-economics guardrails
·Operating model: staffing & RACI across the six stages
·Vendor procurement & third-party model risk workflow
·Quarterly benefits tracking (planned vs realized value)

Future production integrations

·External eval stores
·Observability tool integrations
·Model registry integration
·Vector DB adapter
·Role-based review workflows
·Exportable governance evidence pack

Intentionally out of scope for now

·Real enterprise data connectors
·User authentication
·Full MLOps platform replacement
·Full model-training framework
·Deep PyTorch / TensorFlow notebooks
·Confidential client data
·Cloud infrastructure provisioning

This roadmap keeps the product focused on enterprise AI program delivery rather than turning it into a generic AI course or a full production platform. It intentionally does not implement a transformer or ship training notebooks — the goal is to demonstrate enterprise AI delivery, not to become a deep learning course.

Simulation boundary

This portfolio demo uses deterministic client-side engines and sample data to show enterprise AI delivery mechanics without confidential data or cloud infrastructure. Some signals — production telemetry, incident and regression history — are modeled. The architecture is built around handoff contracts so real integrations can replace modeled signals later.

Implemented / deterministic

●Lifecycle state & handoff contracts
●Strategy scoring & build-path recommendation
●Data readiness derivation
●RAG lab logic (chunking, retrieval, evidence, eval views)
●Operate evidence engine (readiness, lineage, monitoring, regression, incidents)
●Governance decision engine
●Realize ROI engine (leakage, risk discount, payback, NPV)

Modeled / simulated

○Production telemetry
○Real incident history
○Observability-tool integrations
○Real vector database / ANN retrieval
○Real user-feedback stream
○Enterprise data connectors
○Eval run-over-run history