AI Customer Care LLM Setup
Budget / Salary€5,000–10,000
TypeFreelance project
LocationRemote
Posted1 hour ago
I have a working GitHub repository that already contains the base code for a Large Language Model–driven customer-care assistant. Your first task will be to review that repo so you can give an accurate quote; without seeing it, any price discussion would be meaningless.
Project objectives
• Use the existing code to automate responses, analyse incoming customer queries and generate clear, detailed reports.
• Keep the footprint light—our demo will run on a single free GPU instance, so every optimisation that prevents capacity spikes matters.
• Prepare for production by planning scale-out options that I can turn on later, once licence purchases are approved.
Scope of work
1. Repository walkthrough and environment setup on the 1-GPU demo box (Docker or similar).
2. Model fine-tuning or prompt-engineering to hit performance targets without blowing up compute costs.
3. API or webhook integration with my current ticketing/live-chat stack.
4. Reporting module that exposes query analytics and response metrics (simple dashboard or exportable CSV is fine for now).
5. Documentation plus user training so my support staff can run tests and interpret reports.
6. Ongoing technical support, regular updates and maintenance once we move past the demo phase.
Key expectations
• Technical support: quick turnaround on issues during and after deployment.
• User training: short remote sessions and concise manuals.
• Regular updates & maintenance: monthly patching, model refreshes, dependency checks.
Licences for commercial models or plug-ins will be my responsibility; you can work with open-source equivalents in the demo until that paperwork clears.
If you have proven experience with Python, Hugging Face, LangChain (or similar frameworks) and can demonstrate past LLM optimisation on constrained hardware, let’s talk. I will share the private repo as soon as you sign an NDA so you can prepare a precise milestone-based proposal.
Project objectives
• Use the existing code to automate responses, analyse incoming customer queries and generate clear, detailed reports.
• Keep the footprint light—our demo will run on a single free GPU instance, so every optimisation that prevents capacity spikes matters.
• Prepare for production by planning scale-out options that I can turn on later, once licence purchases are approved.
Scope of work
1. Repository walkthrough and environment setup on the 1-GPU demo box (Docker or similar).
2. Model fine-tuning or prompt-engineering to hit performance targets without blowing up compute costs.
3. API or webhook integration with my current ticketing/live-chat stack.
4. Reporting module that exposes query analytics and response metrics (simple dashboard or exportable CSV is fine for now).
5. Documentation plus user training so my support staff can run tests and interpret reports.
6. Ongoing technical support, regular updates and maintenance once we move past the demo phase.
Key expectations
• Technical support: quick turnaround on issues during and after deployment.
• User training: short remote sessions and concise manuals.
• Regular updates & maintenance: monthly patching, model refreshes, dependency checks.
Licences for commercial models or plug-ins will be my responsibility; you can work with open-source equivalents in the demo until that paperwork clears.
If you have proven experience with Python, Hugging Face, LangChain (or similar frameworks) and can demonstrate past LLM optimisation on constrained hardware, let’s talk. I will share the private repo as soon as you sign an NDA so you can prepare a precise milestone-based proposal.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.