The conversations getting HuskyBot off the ground.

Hello RunPod Research Team,

I am writing on behalf of the St. Cloud State University Library to apply for the RunPod Research Credits program. We are building HuskyBot, an open-source academic AI assistant: a custom-trained, fine-tuned model that combines a library chatbot with a research-paper writing assistant.

Project summary:
  • Institution: St. Cloud State University Library
  • Approach: LoRA / QLoRA fine-tuning of an open-weight base model (Mistral 7B or Llama 3 8B)
  • Estimated compute: ~150–250 GPU hours on A100 80GB or H100
  • Timeline: May 15 – August 15, 2026
  • Output: Open-source model + downloadable GGUF + web app

This project will be released as open source under the SCSU Library, and we plan to publish a model card on Hugging Face along with our training methodology. We would appreciate the opportunity to discuss research credit allocation that would help us complete training within our timeline.

Happy to share our full proposal, fine-tuning plan, or any other details that would be useful.

Best regards,
Gustavo Valdez Paez
Lead Developer & Project Coordinator
HuskyBot — SCSU Library AI Initiative

Hello Lambda Labs Team,

I am reaching out from St. Cloud State University Library regarding the Lambda Labs academic discount program (50% off). We are building HuskyBot, an open-source academic AI assistant for our library and research community.

Project details:
  • Institution: SCSU Library (St. Cloud State University, Minnesota)
  • Goal: Fine-tune an open-weight LLM for academic writing assistance and library reference
  • Compute needs: ~150–250 GPU hours on A100 80GB or H100
  • Timeline: Summer 2026 (May 15 – August 15)

We would like to apply the academic discount to our compute reservation. Could you confirm the application process and what documentation you require to verify our academic affiliation?

Thank you for supporting academic research.

Best,
Gustavo Valdez Paez
Lead Developer & Project Coordinator
HuskyBot — SCSU Library

Hello Semantic Scholar Team,

We are reaching out from St. Cloud State University Library regarding bulk access to S2ORC (Semantic Scholar Open Research Corpus) for an open-source academic AI project.

Project overview:
  • Institution: SCSU Library (St. Cloud State University)
  • Project: HuskyBot — open-source academic AI assistant
  • Use case: Training corpus for fine-tuning an open-weight LLM on multidisciplinary academic literature
  • Output: Open-source model published on Hugging Face
  • Timeline: May – August 2026

Could you advise on the appropriate process for obtaining S2ORC bulk access for this purpose, and confirm any usage requirements or attribution we should follow?

Thank you for the incredible work you do supporting open scholarship.

Best regards,
Gustavo Valdez Paez
Lead Developer & Project Coordinator
HuskyBot — SCSU Library

Hello ACM Digital Library Team,

I am writing on behalf of the St. Cloud State University Library to inquire about Text and Data Mining (TDM) licensing options for academic LLM training.

Project context:
  • Institution: SCSU Library (academic, non-commercial use)
  • Project: Open-source academic AI assistant
  • Use: Fine-tuning a small open-weight LLM with academic content
  • Output: Open-source model with attribution to source datasets

We understand that standard TDM agreements often restrict use for LLM training, and we would like to discuss whether ACM offers (or would consider) a negotiated addendum for academic, non-commercial training use — with appropriate attribution and any restrictions you would require.

We would welcome the opportunity to discuss your terms, pricing tiers, and any documentation we should provide to support such a request.

Thank you for your time.

Best regards,
Gustavo Valdez Paez
Lead Developer & Project Coordinator
HuskyBot — SCSU Library

Hello IEEE Xplore Licensing Team,

The St. Cloud State University Library is exploring TDM API access for academic LLM training as part of our HuskyBot project — an open-source academic AI assistant.

Key details:
  • Use case: Fine-tuning a small open-weight LLM (Mistral 7B / Llama 3 8B class) with academic STEM literature
  • Distribution: Open source, with proper attribution to all source datasets
  • Institutional: Non-commercial, SCSU Library
  • Timeline: Phase 2 of our project (June–August 2026)

Could you share information about your TDM API tiers, LLM-training-compatible licensing options, and what would be required from us to move forward?

Thank you for considering this request.

Best,
Gustavo Valdez Paez
Lead Developer & Project Coordinator
HuskyBot — SCSU Library

Hello Elsevier Licensing Team,

We are reaching out from St. Cloud State University Library regarding ScienceDirect TDM API access for an open-source academic AI initiative.

Project brief:
  • SCSU Library — academic, non-commercial
  • Open-source academic chatbot + research writing assistant
  • Fine-tuning an open-weight LLM
  • Open release with full attribution

We understand that LLM training is typically not covered under standard TDM terms. Could you advise on (1) whether a negotiated addendum is possible for academic, non-commercial fine-tuning, and (2) what your evaluation process and timeline look like?

Thank you, and we appreciate your guidance.

Best regards,
Gustavo Valdez Paez
Lead Developer & Project Coordinator
HuskyBot — SCSU Library

Hi [Name],

I hope this message finds you well. I am writing to brief you on HuskyBot — the SCSU Library AI Initiative — and to start a conversation about library systems integration as we approach Phase 3 of the project.

The short version:
  • HuskyBot is an open-source academic AI assistant being built this summer at SCSU Library.
  • It will be available both as a web app embedded in the library website and as a downloadable model for offline use.
  • In Phase 3 (mid-July through August), we plan to integrate with library catalogs, research guides, and reference resources — ideally read-only, lightweight integrations.

What I would like to discuss:
  • What APIs or read-only endpoints are available for our library catalog and discovery layer?
  • Any preferred integration patterns or constraints (auth, rate limits, hosting)?
  • Whether there is interest in eventually embedding the chatbot directly on the library website

Would you have 30 minutes in the next two weeks for a quick conversation? Happy to walk through the project, share our roadmap, and answer any questions.

Thank you for your time — looking forward to it.

Best,
Gustavo Valdez Paez
Lead Developer & Project Coordinator
HuskyBot — SCSU Library

Dear [Dean Name],

I hope your semester is wrapping up well. I wanted to formally introduce HuskyBot — the SCSU Library AI Initiative — and request a brief stakeholder briefing.

Project overview:
  • HuskyBot is an open-source academic AI assistant being developed this summer.
  • It combines a library chatbot with a research-paper writing assistant.
  • The model will be available both as a web app and as a downloadable open-source model.
  • The project is intentionally lean and built for scale.

Our goal:
  Position SCSU Library as a leader in open-source academic AI — an institution that not only adopts AI tools but builds and shares them with the broader community.

Would you be available for a brief 30-minute meeting in the next 2–3 weeks? I would like to walk you through the project plan, discuss alignment with library priorities, and answer any questions.

Thank you for your support of student-led research initiatives.

With respect,
Gustavo Valdez Paez
Lead Developer & Project Coordinator
HuskyBot — SCSU Library AI Initiative