IdeaMVPScale

60+ Products launched since 2018
Deep expertise in building Apps, AI, and Web3
DeFi
Fintech
AdTech
AgTech
HealthTech

Our Clients

Startups From

Case Studies

# Web App

Lend and borrow against income, the new collateral

#dApp  #DeFi  

57Blocks goes beyond just product development to actively contribute to the growth and success of early-stage startups like Huma. This case study explores the four pillars of 57blocks unique and multifaceted partnership with one of the foundational protocols in the decentralized finance (DeFi) space.

# Mobile App

Tokenpad - Smarter crypto investors. Now supporting 14 chains, hundreds of DeFi projects, and thousands of tokens, Tokenpad is a top choice for tracking crypto investments.
Tokenpad - Smarter crypto investors.

Articles

Image Search with AI Models

By using just an image on a mobile device, the search application is designed to return matches from the database that are either identical or bear a high resemblance to the original uploaded image. Here, we describe the technology behind this powerful functionality.

Alpha Xiang
Yanqi Liu
Anjing Wang
Alpha Xiang & Yanqi Liu & Anjing Wang

What Makes a Good Vector Database? Comparing Pinecone and LanceDB

Which vector database is best to use? It depends. Based on our experiences using them, there is no one-size-fits-all "best" database. Rather, a Superior Vector Database is one that is well-matched with its use case, meeting the unique requirements of that scenario. Read our comparative analysis to discover why.

Alex Hu
Steven Zhang
Anjing Wang
Alex Hu & Steven Zhang & Anjing Wang

How to Use LLMs to Extract Document Information

Today, there is no easy way to convert raw text from volumes of academic papers into structured data, or complete information extraction (IE). Traditional IE methods often rely on labor-intensive handcrafted rules and patterns, which may not generalize across diverse domains and languages. We used various Large Language Models (LLMs) from GPT to Claude to complete IE from these documents and compared their performance. We share what we learned here.

Carvin Li
Anjing Wang
Carvin Li & Anjing Wang

Image Search with AI Models

By using just an image on a mobile device, the search application is designed to return matches from the database that are either identical or bear a high resemblance to the original uploaded image. Here, we describe the technology behind this powerful functionality.

Alpha Xiang
Yanqi Liu
Anjing Wang
Alpha Xiang & Yanqi Liu & Anjing Wang

What Makes a Good Vector Database? Comparing Pinecone and LanceDB

Which vector database is best to use? It depends. Based on our experiences using them, there is no one-size-fits-all "best" database. Rather, a Superior Vector Database is one that is well-matched with its use case, meeting the unique requirements of that scenario. Read our comparative analysis to discover why.

Alex Hu
Steven Zhang
Anjing Wang
Alex Hu & Steven Zhang & Anjing Wang

How to Use LLMs to Extract Document Information

Today, there is no easy way to convert raw text from volumes of academic papers into structured data, or complete information extraction (IE). Traditional IE methods often rely on labor-intensive handcrafted rules and patterns, which may not generalize across diverse domains and languages. We used various Large Language Models (LLMs) from GPT to Claude to complete IE from these documents and compared their performance. We share what we learned here.

Carvin Li
Anjing Wang
Carvin Li & Anjing Wang
View Moremore

Pricing

Baseline

PRD, Wireframes, Budget
1-2 weeks

Free

Baseline

PRD, Wireframes, Budget, 1-2 weeks

Free

Show Details

Poc/Mvp

Tech, Product
3-12 weeks

$50k-$100k

Poc/Mvp

Tech, Product, 3-12 weeks

$50k-$100k

Show Details

Scalling Engineering

Extend your budget by 2-3x
dedicated team

Contact Us

Scalling Engineering

Extend your budget by 2-3x
dedicated team

Contact Us

Show Details
About Us

We’ve been finding ways to build better globally distributed engineering teams since 2003.

Our core team has been working together since then and has worked together through 5 different companies now. Together we’ve worked in more than 57 pre-seed startups globally, while working with some of the largest companies in the world to co-create new products at scale.

21
Years
57+
Startups
150+
Team members