We’re excited to launch Fastino’s Zero-Shot Model API, enabling Named Entity Recognition (NER), Personally Identifiable Information (PII) detection, and Function Calling—all without requiring any prior training.

‍

With zero-shot capabilities, you can integrate powerful language understanding into your applications instantly, ensuring high accuracy, security, and efficiency for enterprise-grade AI workflows.

‍

🔗 Get Started with the API

INTRODUCING FMT

‍

We’re introducing Fastino Model Tooling (FMT)—a powerful suite designed to help enterprises finetune language models for custom, agentic, and high-performance tasks.

With FMT, you gain unparalleled control over your models, enabling you to optimize accuracy, efficiency, and task-specific performance without the overhead of traditional finetuning workflows.

‍

FASTINO DOCS ARE LIVE

‍

We’re excited to announce the release of Fastino’s documentation! You can now start using our small models API by exploring our docs and making your first calls today.

Our small models are designed for speed, accuracy, and efficiency, allowing developers to integrate powerful task-optimized LLMs into their workflows with ease.

🔗 Read the Docs and Start Building

Whether you’re experimenting with AI for the first time or integrating Fastino into production, our documentation will help you get up and running quickly.

Check it out now and start calling our small models via the API!

ABSTRACT

‍

Urchade Zaratiana, Ashley Lewis, Dan Iter, Julia White, Oliver Boyd, Riley Carlson @ Fastino AI

‍

We introduce fastino pii, a high-performance model for redacting PersonallyIdentifiable Information (PII) from unstructured text. The model is capable ofidentifying and redacting multiple types of PII, including names, phone numbers,addresses, emails, IP addresses, usernames and other sensitive data. Leveraging theFastino architecture, fastino pii is the new state-of-the-art in F1 scores,outperforming larger and smaller LLMs, including gpt-4o. This report outlinesthe task format and evaluation of fastino pii.

‍

MODEL DESIGN

‍

The fastino pii model leverages the same lightweight architecture as Fastino, optimized for low-latency and high-efficiency processing.

Key features include:

• Training: Trained on diverse datasets containing various forms of PII across multiple domains and languages.

• Lightweight Inference: Optimized for CPU environments to enable cost-effective deployment.

‍

RESULTS

‍

The performance of Fastino PII was evaluated against leading models, including GPT-4o-mini, GPT-4o, and Gemini-1.5-Flash, based on F1 score (accuracy in identifying personally identifiable information) and latency (processing speed).

Among all models tested, Fastino PII achieved the highest F1 score of 96.94, outperforming GPT-4o (96.30), Gemini-1.5-Flash (95.11), and GPT-4o-mini (89.50). In addition to its superior accuracy, Fastino PII demonstrated a substantial advantage in processing speed, with a latency of just 257 milliseconds when running on a CPU. In contrast, the other models exhibited significantly higher latencies: GPT-4o-mini (2812 ms), GPT-4o (2855 ms), and Gemini-1.5-Flash (2450 ms).

This combination of high accuracy and exceptional speed makes Fastino PII an efficient solution for PII redaction. Its high recall ensures that sensitive entities are not overlooked, while its high precision minimizes false positives. These strengths enable Fastino PII to provide fast and reliable PII detection without sacrificing accuracy, even when deployed on limited computational resources.

‍

We’re thrilled to step out of stealth and officially introduce Fastino to the world. We are also excited to announce our $7M pre-seed funding round led by M12, Microsoft’s Venture Fund, and Insight Partners.

‍

OUR MISSION

‍

At Fastino, our mission is simple: enable developers everywhere to do more with AI. We are building a family of developer-first LLMs that are more accurate, faster, and safer.

Fastino was born out of frustration—as developers and AI practitioners, we’ve struggled with:

• Opaque APIs

• The constant race for GPU resources

• Generalized models that require endless prompting

• Unpredictable results and lack of cost control

We knew there had to be a better way.

‍

A NEW ERA OF AI DEVELOPERS

‍

This is why we’re building Fastino—a family of task-optimized language models and tools designed for enterprise flexibility and control.

• Industry-leading accuracy for tasks like:

• Question answering (RAG)

• Textual data structuring

• Text/document summarization

• Smart replies

• Up to 1000x faster than traditional models

• Lower compute requirements

• Deploy anywhere—from commodity CPUs to your virtual private cloud

• Intuitive developer tools for fine-tuning & validation

• Consistent and predictable costs

‍

WORLD CLASS TEAM

‍

To bring this vision to life, we’ve assembled an exceptional team led by our co-founders Ash Lewis and George Hurn-Maloney, alongside top AI researchers and developers from Stanford, Oxford, Berkeley, Microsoft, and Google. We share a commitment to pushing boundaries and delivering real value in AI.

‍

BACKED BY

‍

With the support of world-class investors, we’re just getting started.

Fastino’s $7M pre-seed round was led by Microsoft and Insight Partners, with additional backing from NEA, CRV, Valor, and notable angels including:

• Thomas Dohmke, CEO of GitHub

• Executives from Google

“I’m excited to be an early investor in Fastino, a company on a mission to bring the world accurate, fast, and safe task-specific LLMs that solve organizations’ most pressing challenges. Their novel approach involves a new architecture that runs on CPUs, making AI more accessible for a future with 1B developers.”

— Thomas Dohmke, CEO of GitHub

‍

JOIN US

‍

Want to learn more?

• Join the waitlist to be part of our private alpha and see Fastino in action.

• Explore career opportunities to join our team and help shape the future of enterprise LLMs.

The future of developer-first AI starts now. Let’s build it together.

‍

JOIN FASTINO

‍

At Fastino, we’re building the next generation of LLMs using a novel architecture. Our team is developing high-performance language models that empower enterprise AI teams to achieve more ambitious goals with greater accuracy, speed, and controllability.

We are a challenge-seeking team of former founders, competitive programmers, and leaders from Stanford, Berkeley, Microsoft, Google, and other cutting-edge institutions and companies at the forefront of AI.

Building our family of models is just the beginning. We are committed to making training and deploying models easier for developers—on any planet. If you’re passionate about tackling engineering challenges that will redefine what’s possible in enterprise AI, we invite you to join us.

‍

NOW HIRING

‍

We are currently accepting applications for the following roles:

• Research Scientist

• Machine Learning Engineer

• Software Engineer

• Head of Talent

• General Applications

‍

To apply, send your resume to jobs@fastino.ai.

Help us shape the future of enterprise AI. Let’s build something extraordinary together.

‍