You will get a natural language chat interface for your data
Project details
I have built multiple multi data source agent harnesses that poll the database based on the users question. Getting the AI to pull together information from multiple sources and keep it in context is a real dance. The real key is sanitizing the data before it ever hits the model. Garbage in, garbage out is twice as true with LLMs. I heavily recommend Anthropic's Claude. Its 1m context model has worked amazingly well with some of the larger datasets I have worked with.
The ISP tool I built for a client is a good example. They needed to query FCC broadband availability data across thousands of US towns. I downloaded the bulk CSV files, indexed them with DuckDB, and built a chat interface where they type a town name and get back a structured market report. No SQL, no dashboards, just questions and answers. I can do the same thing for your data whether it is a spreadsheet, a database, a document collection, or an API.
The ISP tool I built for a client is a good example. They needed to query FCC broadband availability data across thousands of US towns. I downloaded the bulk CSV files, indexed them with DuckDB, and built a chat interface where they type a town name and get back a structured market report. No SQL, no dashboards, just questions and answers. I can do the same thing for your data whether it is a spreadsheet, a database, a document collection, or an API.
Programming Languages
HTML & CSS, TypeScriptWhat's included
| Service Tiers |
Starter
$240
|
Standard
$300
|
Advanced
$500
|
|---|---|---|---|
| Delivery Time | 5 days | 5 days | 10 days |
Number of Revisions | 9 | 9 | 9 |
Design Customization | - | - | - |
Content Upload | - | - | - |
Responsive Design | - | - | - |
Source Code | - | - | - |
About Corey
AI-Driven Software Engineer | Systems, ML & Full-Stack
Ava, United States - 8:26 pm local time
am good at getting real results, not demos.
I have reimplemented PyTorch op-for-op in pure C, built a ternary
training optimizer running native {-1, 0, +1} with no FP32 shadow
weights, written a UEFI inference engine that boots straight to
ring 0 with no OS underneath it, and shipped custom Lua firmware
for an ESP32 based ereader. Full stack apps in
Quasar/Vue/TypeScript too.
I care about outcomes. Systems, ML, full apps, it doesn't matter.
I produce complete working results. The repos are public and the
work speaks for itself.
Steps for completing your project
After purchasing the project, send requirements so Corey can start the project.
Delivery time starts when Corey receives requirements from you.
Corey works on your project following the steps below.
Revisions may occur after the delivery date.
reach out and give me your scope.
how many data sources and what kind of data we will be working with.
if we agree your project fits
I will give you a list of everything I need from you