Custom GPT Chatbot for Product Matching -- 2
Budget: $30 – $250 USD
I'm seeking a professional who can develop a custom GPT-enabled chatbot, tailored for a product description matching app.
Key Tasks:
- The chatbot's core purpose will be to perform keyword analysis and category matching, which means it must be capable of scrutinizing and comparing product descriptions effectively.
Crucial Skills and Experience:
- Custom GPT based on existing commercial offers
- Natural Language Processing (NLP)
- Keyword Analysis
- Category Matching
The right freelancer for this project will have a strong understanding of chatbot development and relevant experience with GPT models. Experience in using these technologies for product matching purposes will be highly advantageous. Demonstration of past successful projects of similar nature would also be appealing.
Main function: bot serves to matches 'client's product descriptions' to the company's product names. The user will then reply if the product name is 'correct, incorrect, not sure'. All replies must feed back the dataset and the bot for incremental learning (machine learning with human interaction).
Scenario: our clients name the products based on their descriptions and not on their names. For instance, for a "Diet Coke" they name: "cola based soda with no sugar"; or "sugarless cola based soda"; or "cola soda, no sugar, fizzy drink"; etc. Combinations are endless. Our bot must identify all these possible descriptions and name the product to how we name it "Diet Coke".
Dataset: Attached with dummy products (I'll share the official one after awarding). It's 735 entries two-column dataset where the first column contains the 'company's products names'. The second column contains many different descriptions. For descriptions associated with 'n/a' in the dataset, indicating the absence of a corresponding product in our portfolio.
Language: the dataset is in spanish.
GUI: final POV most have graphic interface -- not command line. Web page must contain a text box and hit button. We'll simply paste a product description like "Fórmulas suplementarias pediátricas de 900 grs. aproximadamente" and it should reply with the product name.
Hosting and coding: it should be hosted anywhere in the cloud -- not locally. You can choose any tool you're most familiar with. We want you to find a tool to use, we will pay for hosting if needed. Prices for that must be shared in advance.
Tone:
- not conversational, very straightforward.
- it should have very straightforward replies: only the 'product name', nothing else. In case it doesn't know the name it should say that it 'doesn't have enough data to know the name', in case it has censorable doubts between 2 or more products it should present those two or more names
Key Tasks:
- The chatbot's core purpose will be to perform keyword analysis and category matching, which means it must be capable of scrutinizing and comparing product descriptions effectively.
Crucial Skills and Experience:
- Custom GPT based on existing commercial offers
- Natural Language Processing (NLP)
- Keyword Analysis
- Category Matching
The right freelancer for this project will have a strong understanding of chatbot development and relevant experience with GPT models. Experience in using these technologies for product matching purposes will be highly advantageous. Demonstration of past successful projects of similar nature would also be appealing.
Main function: bot serves to matches 'client's product descriptions' to the company's product names. The user will then reply if the product name is 'correct, incorrect, not sure'. All replies must feed back the dataset and the bot for incremental learning (machine learning with human interaction).
Scenario: our clients name the products based on their descriptions and not on their names. For instance, for a "Diet Coke" they name: "cola based soda with no sugar"; or "sugarless cola based soda"; or "cola soda, no sugar, fizzy drink"; etc. Combinations are endless. Our bot must identify all these possible descriptions and name the product to how we name it "Diet Coke".
Dataset: Attached with dummy products (I'll share the official one after awarding). It's 735 entries two-column dataset where the first column contains the 'company's products names'. The second column contains many different descriptions. For descriptions associated with 'n/a' in the dataset, indicating the absence of a corresponding product in our portfolio.
Language: the dataset is in spanish.
GUI: final POV most have graphic interface -- not command line. Web page must contain a text box and hit button. We'll simply paste a product description like "Fórmulas suplementarias pediátricas de 900 grs. aproximadamente" and it should reply with the product name.
Hosting and coding: it should be hosted anywhere in the cloud -- not locally. You can choose any tool you're most familiar with. We want you to find a tool to use, we will pay for hosting if needed. Prices for that must be shared in advance.
Tone:
- not conversational, very straightforward.
- it should have very straightforward replies: only the 'product name', nothing else. In case it doesn't know the name it should say that it 'doesn't have enough data to know the name', in case it has censorable doubts between 2 or more products it should present those two or more names