Example projectWeb appAIOpen-source model
On-premise report checker
Hypothetical. Not a client, not a result.
// The problem
Supervisors return a large share of reports for missing elements, and officers wait days to find out a report needs a change.
// What we would build
An assistant that checks a report against the service's own report standards before it is submitted, pointing out missing elements and inconsistencies for the officer to fix. It never writes the narrative. It runs on an open-source model on a GPU server inside the service, so no report leaves the building.
// What is in it
- Your report standards and policies indexed as the reference
- Missing elements and inconsistencies pointed out before submission
- The officer writes and changes the report; the assistant only comments
- An open-source model on a GPU server on premises, with no outside calls
- An evaluation set from past returned reports, run before every change
// Stack
- Llama or Mistral (open-source)
- vLLM
- PostgreSQL with pgvector
- Next.js
- Lenovo GPU server on premises
// estimate
- Build
- ≈ US$112,000 to US$171,000, delivered within 22 weeksCAD 159,800 to 242,800
- 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 3.9 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
- 0.4 to 0.9 weeks
- Penetration test findings closed
- 0.9 to 1.4 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
- Independent penetration test and retest
- 2 to 4 weeks
How it is paid
- Deposit 20%
- CAD 31,900 to 48,500
- Discovery 1.7%
- CAD 2,711.50 to 4,122.50
- Specification and evaluation plan 9%
- CAD 14,355 to 21,825
- Design approved 1.7%
- CAD 2,711.50 to 4,122.50
- Core features 5.2%
- CAD 8,294 to 12,610
- Full build 3.5%
- CAD 5,582.50 to 8,487.50
- Working pilot 27.2%
- CAD 43,384 to 65,960
- Testing and fixes 1.7%
- CAD 2,711.50 to 4,122.50
- Penetration test findings closed 8.7%
- CAD 13,876.50 to 21,097.50
- Launch 1.7%
- CAD 2,711.50 to 4,122.50
- Production 9.6%
- CAD 15,312 to 23,280
- Holdback, 30 days after launch (10%)
- CAD 15,950 to 24,250
- 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