We define the test before we recommend the tool.

Neso AGI research notesReviewed September 2026

Neso AGI is the research lab inside AnovaGrowth. These notes record the questions, methods and review gates behind our model, agent and memory recommendations. Concepts are labeled as concepts, and nothing here is presented as a validated client result.

The notes

  • Research track

    AnovaAI LLM

    A research track for evaluating compact models, retrieval, and tool-assisted workflows against clear operating requirements.

    Read the note

  • Concept

    DomeAI

    DomeAI explores persistent, searchable memory for AI agents, with explicit source context, access boundaries, and a reviewable record of what informed each action.

    Read the note

  • Concept

    Advanced Unified Radar Architecture

    Project AURA is a radar-fusion research concept: could several radar observations be aligned in one volumetric view that keeps uncertainty visible? It studies the interface, the data shaping and the review questions behind that idea.

    Read the note

  • Concept

    Weather-aware planning research

    Could machine learning make one specific weather-sensitive planning decision more useful? This concept frames how to find out. It is not an active forecasting model, an accuracy result or a product.

    Read the note

  • Protocol draft

    What we're testing right now

    Research questions, evaluation boundaries and protocols under review. Results are published only when the run data and scoring method are ready to inspect.

    Read the note

  • Protocol draft

    Evaluation methods before conclusions

    Draft methods for testing model performance, agent behavior, efficiency and safety. These briefs define what must be measured before the lab publishes a result.

    Read the note

  • Concept

    Research concepts and review gates

    Ideas the lab may investigate, not announced products or active previews. Each one needs a scoped feasibility review before any prototype or outside evaluation.

    Read the note

  • Editorial

    Progress to AGI

    Our honest, opinionated read on how close AI is to artificial general intelligence. It changes as the field does. Take it for what it is: one team's informed perspective.

    Read the note

What the lab is looking at

Questions across models, retrieval, hosting and workflows. Each one is a question, not a finding.

  • Model distillation: compact-model tradeoffs across task quality, latency and cost.
  • Model training: the data, holdout tests and review gates that domain adaptation needs.
  • Image and video: provenance, consistency and review for generated media.
  • Evaluation tracking: public evaluations compared against declared models, tasks and conditions.
  • Agentic systems: how tool-using agents prepare actions while keeping approval boundaries.
  • Chat and support: source-grounded conversations for support, sales and internal operations.
  • Small language models: compact local models against larger references on specific tasks.
  • Hosted models: hosting, privacy, observability and throughput tradeoffs.
  • Task models: narrow models for repeatable workflow steps with clear failure handling.

From question to evidence

Every proposed experiment defines its measures before results are shown. The method records task quality, reliability, latency and review load, so a recommendation can be judged against a real operating need.

  1. Research: explore the architectures, training methods and model behaviors that matter for a real task.
  2. Prototype: small, controlled trials to test a hypothesis quickly.
  3. Review: a method moves forward only after its evidence and approval gates are defined.
  4. Document: methods, operating conditions and limitations are written down once the evidence is ready.

Questions about the lab.

What is Neso AGI?

Neso AGI, the research lab inside AnovaGrowth, organizes research questions and evaluation methods across models, retrieval, tool use, and reviewed workflows. Concept pages are not validated products or published benchmarks.

Can I use the models or tools you build in the labs?

There is no standing public model catalog or preview program represented here. Contact us if you have a concrete task, dataset, or risk question that could support a scoped feasibility review.

How does the labs work relate to your client services?

The labs pages define questions and review gates that can inform client recommendations. A method should only be recommended when its evidence matches the client task and operating conditions.

Is the weather prediction research available to use?

No finished weather model or comparative accuracy benchmark is represented. The page describes questions a future evaluation would need to answer across regions, conditions, baselines, and uncertainty.

What does "Progress to AGI" mean?

It is our subjective, internal assessment of how close current AI systems are to artificial general intelligence. It is not a scientific measurement. Think of it as our team's informed perspective on where the technology stands today.

How can I collaborate with the lab?

Use the contact page to propose a specific research question, dataset contribution, or protocol review. Feasibility and scope are confirmed case by case.

Research with a review boundary.

Protocols, evidence and limits get written down before a method is recommended for real use. Bring a concrete task and we will define what has to be proven.

Bring us a taskNeso AGI