Problem Description

We want to build an internal UI that is reminiscent of LinkedIn or golden.com. It will be used to surface data insights across many sources - the goal being to search for people and explore their profiles from there. We expect the ideal front search page could be built using LlamaIndex (GPTIndex) or ElasticSearch.

We have design docs and lots more details on the ask that we can discuss under NDA.

Acceptance Criteria

We expect this project to be iterative so we are looking for a good partner to us as we not only build the MVP but also do a series of improvements over time.

• We want the UI to be clean, logical, and fast for users.
• We expect to see several types of pages (1) home page with semantic search and (2) alternative option for advanced search, (3) search results page, and (4+) people/company/other entity profile pages.
• We want the search results to be powerful to understand quickly with rankings and insights into why those rankings. We would also like the ability to dynamically change weightings that go into the rankings of the search results.
• We want this to be built in such a way that we can continuously and flexibly improve upon it.
• You can expect the data to come to the surface pre-cleaned and pre-processed, so it is not necessary to have ETL or data engineering experience. But if you do have that type of experience or network graph experience and want to explore that path, that's cool too.

More Details on the Specifics

• The home page will feature a chatGPT style search bar that takes in semantic text. Probably using LlamaIndex or ElasticSearch.
• The search can also be done with "advanced search" -> which would essentially surface the criteria to do no code SQL search.
• The results of that search will be a prioritized list of profiles which show why they are ranked in that order.
• Clicking into any profile page will surface a baseball card style page about the person, surfacing lots of insights from many data sources that we have, key features, ways to engage etc.
• There will also be profile style pages for other entities that one might see on a person's page (think of this like clicking on the company page from someone's LinkedIn profile page).
• As we look towards future iterations of this, we are hoping to add fun visuals working on network connections of the people, recommendations for similar people, better semantic understanding of what people are like, and all sorts of other fun features (we are open to being creative here).

Technical Details

• We are flexible on the language built here but would like it to be something that someone else (that is reasonably technical) can pop into - something like Flask, Django, or React.
• We need to make a decision on what database/API should be used for final layer ingestion of data to the UI. It will originally be coming from a cloud warehouse.

Timelines / Milestones

• 3/31 - have partner selected and had deep design sessions
• 4/10 - have first MVP of product ready for testing
• 4/15 - MVP is ready to go
• Post 4/15 - continuous improvements (priced as appropriate for future asks)