Name: Belitsoft
Telephone: +1 (917) 410-57-57

Large Language Models blog

BloombergGPT is Live. A Custom Large Language Model for Finance

Bloomberg is a financial technology and data company with a core business - a subscription-based software platform (Terminal) used by finance professionals for real-time data, analytics, trading tools, and news (this is what Wall Street traders, analysts, and portfolio managers rely on daily). Bloomberg has a strong AI and machine learning division, which developed BloombergGPT, its own financial large language model trained on proprietary financial data. So, it’s time to take the lesson, if you want to build or modernize your own financial product based on a custom LLM and do it smarter, faster, and cheaper.

Dzmitry Garbar • 8 min read

Financial LLM: Use Cases and Examples

A major expectation in banking is that LLMs can serve as intelligent assistants for both customers and employees. In the world of asset management, including hedge funds, mutual funds, and other investment firms, the interest in financial LLMs centers on gaining an informational edge and productivity boost. Fintech companies and startups view financial LLMs as an opportunity to differentiate their products with AI and to build new services faster. When the insurance industry speaks of a financial LLM (sometimes explicitly an "insurance LLM"), they mean a model that can understand insurance-specific language and workflows. Companies that provide financial data, analytics, and news have been quick to explore LLMs, often coining their solutions as financial LLMs to market their domain expertise. Their perspective is that a financial LLM should function as an expert financial analyst that a user can talk to on demand.

Dzmitry Garbar • 4 min read

Hugging Face: The Guide for AI Startup Founders

Hugging Face (HF) is a leading open-source platform and community in the machine learning ecosystem. As the “GitHub of Machine Learning”, HF is a central place to find, share, and collaborate on AI models and datasets. HF is a one-stop resource in the AI development pipeline, from model search and experimentation to initial deployment, helping startups go from an idea to a working AI faster and with lower upfront R&D costs.

Dmitry Baraishuk • 11 min read

Cost of Building Text Classification Model for AI-based Content Curation

Our clients often ask us about the cost of building an AI document classification software that automatically analyzes, scores, selects, and prepares large volumes of content for future business use.

Alexander Kom • 7 min read

Truewind AI ($17M): AI Use Case in Finance and Accounting

Generative AI, machine learning capabilities, and predictive analytics are gradually becoming part of modern applications. The finance and accounting industry is no exception: startups are emerging. They not only attract investment but also have real clients. One of them is Truewind — founded just around 2 years ago (at the time of writing). A core component of its offer for companies working in accounting is document classification, assigning categories, displaying a confidence score for classification accuracy, and explaining why a particular classification was applied. Integrating such a solution into a B2B SaaS product enables building a sustainable business model, gradually scaling and deepening automation in accounting processes. Truewind relies less on custom large language models, and its senior full-stack engineers build the application using the OpenAI API and vector databases.

Dmitry Baraishuk • 10 min read

Generative AI App Development Company

This article is a full roadmap for companies looking to build serious generative AI products, showing the process, potential obstacles, and how Belitsoft helps clients go from proof-of-concept to production — with the right team, clear planning, and business-focused execution.

Dmitry Baraishuk • 7 min read

Custom LLM Case Study: Healthcare (Innovaccer, Unicorn)

Custom LLM Case Study: Innovaccer (Healthcare)

SaaS enterprises adopt custom Gen AI. Innovaccer is an example of an enterprise SaaS company that chose a product diversification strategy by introducing its own proprietary LLM models instead of using ChatGPT. They have been training the model on data prepared by Innovaccer’s team and hosting it on servers controlled by Innovaccer. By 2027, most GenAI models used by enterprises will be industry- or function-specific, predicts Gartner. This means an increasing number of custom LLMs will be trained on proprietary data and hosted on servers controlled by enterprises.

Dmitry Baraishuk • 5 min read

Let's Talk Business

Do you have a software development project to implement? We have people to work on it. We will be glad to answer all your questions as well as estimate any project of yours. Use the form below to describe the project and we will get in touch with you within 1 business day.

Contact form

Full name *

E-mail *

Country *

Phone number

Message

Choose File

We will process your personal data as described in the privacy notice

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply

Call us

USA +1 (917) 410-57-57

UK +44 (20) 3318-18-53

Email us

[email protected]