Example projectWeb appAIOpen-source model
Private funder reporting
Hypothetical. Not a client, not a result.
// The problem
Program staff spend days each quarter rebuilding the same participation and spending figures in a different template for each funder.
// What we would build
A reporting workspace that builds each funder's figures from program records, and an assistant on an open-source model running on a server the Nation owns that prepares narrative sections from the Nation's own records for staff to edit and approve. No community data leaves the Nation's server.
// What is in it
- Participation and spending figures built from program records
- Templates for each funder, editable by staff
- Narrative sections prepared from your own records, edited and approved by staff
- An open-source model on a GPU server owned by the Nation
- An evaluation set from past reports, run before every change
// Stack
- Llama or Mistral (open-source)
- vLLM
- PostgreSQL with pgvector
- Next.js
- Lenovo GPU server owned by the Nation
// estimate
- Build
- ≈ US$112,000 to US$170,000, delivered within 23 weeksCAD 159,500 to 242,400
- Hosting
- ≈ US$176 a monthCAD 250 a month
- Support
- ≈ US$2,500 a monthCAD 3,565 a month
- Hardware
- ≈ US$47,800 to US$197,000CAD 68,000 to 280,800
Prices in your currency are estimates from today's Bank of Canada rate. All invoicing is in CAD or USD.
- AI route
- Open-source model
- Model running cost
- ≈ US$176 a monthCAD 250 a month
Timeline by milestone
- Discovery
- 2.9 to 3.4 weeks
- Specification and evaluation plan
- 0.4 weeks
- Design approved
- 2.4 to 3.4 weeks
- Data governance rules approved by your council or committee
- 1.4 to 2.4 weeks
- Core features
- 1.4 to 1.9 weeks
- Full build
- 0.9 to 1.4 weeks
- Working pilot
- 3.4 to 5.9 weeks
- Testing and fixes
- 0.9 to 1.9 weeks
- Launch
- 0.9 to 1.4 weeks
- Production
- 0.9 to 1.4 weeks
Outside our hands, and added to the calendar
- Hardware delivery, after it is ordered
- 2 to 6 weeks
How it is paid
- Deposit 20%
- CAD 31,840 to 48,420
- Discovery 1.7%
- CAD 2,706.40 to 4,115.70
- Specification and evaluation plan 9%
- CAD 14,328 to 21,789
- Design approved 1.7%
- CAD 2,706.40 to 4,115.70
- Data governance rules approved by your council or committee 8.7%
- CAD 13,850.40 to 21,062.70
- Core features 5.2%
- CAD 8,278.40 to 12,589.20
- Full build 3.5%
- CAD 5,572.00 to 8,473.50
- Working pilot 27.2%
- CAD 43,302.40 to 65,851.20
- Testing and fixes 1.7%
- CAD 2,706.40 to 4,115.70
- Launch 1.7%
- CAD 2,706.40 to 4,115.70
- Production 9.6%
- CAD 15,283.20 to 23,241.60
- Holdback, 30 days after launch (10%)
- CAD 15,920 to 24,210
- GPU time for training and testing, at cost
- CAD 300. Billed up front, at cost, outside the milestones
- GPU server for AI, for the model
- CAD 68,000 to 280,800. Billed up front, at cost, outside the milestones