What Fleak does

Fleak is a private, venture-backed AI infrastructure company founded in 2024 and headquartered in San Jose, California, operating under the legal entity Fleak Tech Inc. The company positions itself as a "Universal Schema Intelligence layer" between enterprise data sources and AI applications, addressing the data quality problem that limits the effectiveness of AI/ML systems. Its core product is the Fleak Platform, built on ZephFlow — a lightweight, embeddable data processing engine that executes DAG-based transformation pipelines in memory. The platform offers value-aware routing (AI-evaluated routing of incoming data points based on downstream application requirements), AI-orchestrated self-healing pipelines (3-minute average recovery from schema drift via automatic config regeneration), governed delivery with full audit trails, and a natural language copilot for workflow creation. Fleak is SOC 2 Type II certified and supports four deployment modes — fully managed SaaS, customer cloud, on-premise, and air-gapped — to address regulated buyer requirements.

The product portfolio spans Fleak Data Workflow (a managed DAG-builder UI), ZephFlow (embeddable as JVM SDK, Docker container, or HTTP API), processing nodes (Parser, Filter, Eval, SQL, Enrichment, PII Mask), and four vertical-specific solution modules — Cybersecurity (SOC/SIEM/SDR), Industrial/OT (manufacturing, energy), Aviation & Logistics (cargo, MRP), and Financial Services (compliance, fraud). The platform normalizes data to standard schemas (OCSF, UDM, CSF) and integrates natively with 50+ systems across SIEM, OT/IoT, cloud (Kafka, S3, GCS, Azure), databases (PostgreSQL JDBC, Delta Lake, Databricks Unity Catalog), and messaging (SQS, Pub/Sub). Co-founders Bo Lei (Netflix data mesh, Splunk UBA) and Yichen Jin (quant fund data infrastructure, healthcare fraud detection) anchor the technical and commercial narrative.

Fleak operates a hybrid product-led/direct-sales GTM, combining self-serve platform access (app.fleak.ai), developer-facing documentation (docs.fleak.ai), and a Slack community (fleak-hq) with founder-led 30-minute demo sales. Pricing is quote-based, annual subscription, with no public price list disclosed. Two named customers are referenced — Crest Data (enterprise) and aaww/Alpha Level (startup cohort peer). The company has raised approximately $7.37M in disclosed funding across an April 2024 seed and an October 2025 round, and was selected for the inaugural Databricks AI Accelerator Program in September 2025.

Name

Fleak

Legal name

Fleak Tech Inc.

Website

https://fleak.ai

Company type

Private

Founded year

2024

Operating status

Operating

Headcount range

11–50 employees

Short description

Fleak is a private AI infrastructure company that normalizes, filters, and governs real-time data flowing between enterprise sources and AI applications, using its ZephFlow engine for self-healing pipelines. It serves cybersecurity, industrial/OT, aviation, and financial services customers.

Ownership category

akta.pro rank

Product category

Real-time Data Pipeline Infrastructure

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Custom Computer Programming Services (541511)

SIC

Services-Computer Integrated Systems Design (7373), Services-Computer Programming, Data Processing, Etc. (7370)

akta.pro primary industry

AI/ML Solution Integration & MLOps Enablement (BPAEAAAJ)

akta.pro secondary industries

Data & Analytics Platforms (Data Warehousing, Lakes, Streaming) (HDABAAAF), AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ)

Keywords

• Real-time data normalization
• AI data pipeline platform
• Schema drift remediation
• SIEM data ingestion
• OT IoT data integration

Headquarters

HQ city

San Jose

HQ country

United States

HQ region

North America

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Infrastructure, Marketing or Sales, Operations

Revenue model

1. Managed Data Workflow Platform: Fleak generates revenue through a managed data workflow platform subscription model. The platform provides AI-powered data pipeline infrastructure with self-healing capabilities, value-aware routing, and governed delivery. Pricing appears to be quote-based for enterprise deployments with deployment options ranging from fully managed SaaS to air-gapped installations.

Pricing tiers

Go-to-market motion1 record

Distribution channels2 records

Marketing channels5 records

Core offering

Fleak sells an AI infrastructure layer that sits between enterprise data sources and AI/ML applications, normalizing, filtering, governing, and delivering data in real time. The platform runs on ZephFlow, an embeddable DAG-based data processing engine, and is delivered as a managed SaaS workflow builder with self-healing pipelines and value-aware routing. Deployments span Fully Managed SaaS, customer cloud, on-premise, and air-gapped environments.

Product overview

Fleak is an AI infrastructure layer company offering a unified platform for real-time data normalization, filtering, governance, and delivery. The core product is the Fleak Platform with two main components: Fleak Data Workflow (a managed DAG-builder UI for building distributed pipelines) and ZephFlow (a lightweight, embeddable data processing engine for JVM/Docker/HTTP deployments). The platform provides specialized solutions for Cybersecurity, Industrial/OT, Aviation & Logistics, and Financial Services verticals. Key capabilities include AI-orchestrated self-healing pipelines, value-aware routing, LLM token optimization (up to 40% reduction), and automatic schema drift detection. The product portfolio includes processing nodes (Parser, Filter, Eval, SQL, Enrichment, PII Mask, No-op) and supports 50+ native source/sink integrations including Kafka, S3, Databricks, Azure services, GCP services, Splunk, Elasticsearch, and various OT/IoT systems.

Differentiator

Problem solved

Functional benefit

Brands

• ZephFlow: A lightweight, embeddable data processing engine that runs DAG-based transformation pipelines in JVM applications, Docker containers, or as an HTTP API service. ZephFlow provides stateless data processing with support for various source connectors, processing nodes, and multiple sink destinations including Kafka, S3, HTTP, databases, and messaging systems.

Products and services

• Fleak Platform Managed AI infrastructure platform that connects to 50+ enterprise data sources, normalizes and filters data in real time, governs downstream delivery, and self-heals schema drift. Sold to enterprise SOC/SIEM, industrial/OT, financial services, and aviation/logistics teams that feed AI/ML and LLM-powered applications.
• Fleak Data Workflow Managed data workflow product with a DAG builder UI for creating, testing, deploying, and monitoring distributed data pipelines. Includes a natural-language copilot that builds workflows from plain-language descriptions. Targets enterprise data and platform engineering teams building ingestion pipelines for AI applications.
• ZephFlow Embeddable data processing engine with a JVM SDK, Docker container, and HTTP API deployment modes that runs stateless DAG-based transformation pipelines with dead-letter queue support and built-in metrics. Designed for developers who need to embed Fleak's pipeline runtime directly into their own applications or run it as a standalone service.

Quantifiable outcome

• 90% integration cost reduction vs traditional pipeline development
• +5 more outcomes

Named customers2 records

Segments4 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration21 records

AI capability5 records

Feature6 records

Partnerships

One partnership is on record.

• DatabricksflagshipStrategic or Co-development Partner · 18 September 2025Fleak was selected for the inaugural Databricks AI Accelerator Program cohort in September 2025. The program offers pre-seed and seed-stage startups up to $250,000 in funding, product credits, professional services credits, mentorship, and access to the Databricks platform. Participants connect to the Databricks VC Network, a coalition of investors including Andreessen Horowitz, Battery Ventures, General Catalyst, and others. Fleak was recognized for applications in enterprise automation, security, observability, and consumer data usability.

Scale indicators8 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Confluent: Confluent's Kafka-based streaming platform is a core component of the same enterprise data-pipeline stack Fleak serves. Fleak explicitly integrates with Kafka and adjacent runtimes (Vector, Redpanda, Benthos) that are core to the Confluent ecosystem, competing for the same budget dollars on streaming data infrastructure.
• Fivetran: Fivetran is a leading managed data integration/ELT platform that competes directly with Fleak's value proposition of automated, low-maintenance data pipeline onboarding. Both target enterprises seeking to reduce time spent on pipeline engineering.
• Airbyte: Airbyte provides open-source data integration with a large connector catalog, directly competing with Fleak's '50+ pre-built integrations' pitch. Both target the same developer/enterprise buyer looking to reduce bespoke connector work.
• Cribl: Cribl focuses on observability/SIEM pipeline routing and cost reduction (e.g., Splunk cost optimization), directly overlapping Fleak's cybersecurity and SIEM cost-optimization solutions. Both sell into SOC teams that need to filter/route security telemetry.
• Redpanda: Redpanda (acquired Redpanda Connect/Benthos) provides a Kafka-compatible streaming platform with built-in data transformation — Fleak explicitly integrates with Benthos as a runtime target. They compete for the same streaming pipeline buyer.

Broad incumbents

• Databricks: Databricks is both a partner (AI Accelerator) and the largest potential competitor. Its Lakehouse Platform, DLT (Delta Live Tables), and Unity Catalog increasingly absorb the kind of ingestion, normalization, and governance work Fleak sells — and it brings a vastly larger customer base and R&D budget.
• Snowflake: Snowflake's data cloud includes ingestion, transformation, and governance capabilities that overlap with Fleak's pipeline automation. As Snowflake pushes into AI-ready data with features like Dynamic Tables and Cortex, it can absorb Fleak-style functionality into broader enterprise deals.

Emerging players

• dbt Labs: dbt (data build tool) is the leading transformation-layer platform, frequently deployed alongside ingestion tooling. While dbt focuses on transformation rather than ingestion/normalization, it competes for the same 'analytics engineering' buyer and roadmap expansion threatens overlap.
• Monte Carlo: Monte Carlo is a data observability platform focused on detecting data quality issues, schema changes, and pipeline failures — overlapping with Fleak's self-healing and schema-drift capabilities. Both target data engineering teams who need reliability in production pipelines.

Others

• Ingest: Apache Beam and the broader Google Cloud Dataflow ecosystem represent infrastructure-level alternatives for real-time streaming pipelines. While not direct competitors in the same product category, they are alternative architectural choices for the same enterprise data problems.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• GitHubgithub.com/fleaktech/zephflow-core
• Xx.com/getonfleak
• LinkedInlinkedin.com/company/fleak

Compliance1 record

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds3 records

Investors18 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Fleak