When a hospital's diagnostic model is updated, or a bank's fraud detection agent is forked, the question of what changed, who approved it, and what evidence supports the new version is not just technical. It is a matter of clinical and financial accountability. Forkit Dev, a 2025-founded startup from Heilbronn, Germany, is betting that the answer lies in a persistent identity layer it calls an AI Passport, a governance abstraction designed to sit atop the existing sprawl of MLOps and model registries [Forkit.dev, 2024].
The Passport as a Governance Abstraction
The company's core proposition is a structured record that follows a model or agent from development through deployment. This Passport connects identity, lineage, runtime signals, and human approvals into a single, queryable artifact [Forkit.dev, 2024]. The goal is to provide the audit trail that regulated industries, from healthcare to finance, will require as AI moves from pilot to patient care or loan approval. Forkit Dev is not building another model hub or training platform. Instead, it positions its infrastructure as a control plane that links to existing tools, aiming to bring clarity to the often-opaque lifecycle of a production AI system [F6S, 2024].
A Dual-Track Go-to-Market
To drive adoption, the company is pursuing a classic open-source and commercial strategy. Developers can start with two entry points.
- Forkit Dev Core. An Apache 2.0 licensed open-source library for generating and validating Passport files locally, integrating with CI/CD pipelines for automated provenance checks [Forkit.dev, 2024].
- Hosted Passport Flow. A free, cloud-based service where researchers and teams can register AI assets to begin building an evidence record without installing software [Forkit.dev, 2024].
The wedge is clear: lower the friction for developers to experiment with provenance, with the long-term aim of converting mid-size and enterprise teams into customers for managed, scalable deployments. The founders, Arpita Sarker and Syed Amir Hamza, are both PhD candidates whose public work focuses on trustworthy systems, a background that aligns with the product's compliance-by-design ethos [LinkedIn, 2024]. The company is incubated at Campus Founders and backed by Google for Startups, early signals of institutional validation for its technical approach [F6S, 2024].
The Team and Early Backing
The founding team brings a research-intensive perspective to an enterprise problem. Arpita Sarker, listed as Founder & CEO, is also a co-founder of the Deep Tech Collective e.V., indicating a focus on ecosystem building [LinkedIn, 2024]. Syed Amir Hamza serves as Co-founder and CTO [LinkedIn, 2024]. Their small team (estimated at 1-10 employees) is targeting a complex, nascent market where demand is driven more by impending regulation than by immediate feature requests [LinkedIn, 2024].
| Role | Name | Background Note |
|---|---|---|
| Founder & CEO | Arpita Sarker | PhD Researcher, Co-Founder of Deep Tech Collective e.V. [LinkedIn, 2024] |
| Co-founder & CTO | Syed Amir Hamza | PhD Candidate, AI Researcher [LinkedIn, 2024] |
Where the Friction Lies
The ambition is significant, but the path is lined with execution risks familiar to any infrastructure startup. The primary challenge is not technical novelty but commercial coordination. Forkit Dev must convince enterprises to adopt a new governance layer that requires integration across often-siloed data science, engineering, and compliance teams. Its success hinges on becoming the default schema for AI lineage, a standard that larger incumbents like cloud providers or model registry vendors could attempt to define themselves. Furthermore, while the open-source core drives developer love, the business model for converting free users into paying enterprise customers remains unproven. The company's undisclosed pre-seed funding suggests it is in the earliest stages of proving this motion [Forkit.dev, 2024].
For patients and consumers, the disease state here is a lack of transparency in automated decision-making. The patient population is anyone subject to an AI-driven outcome in a high-stakes domain: a loan applicant, a medical imaging patient, or a job candidate screened by an algorithm. Today, the standard of care is fragmented. Evidence is scattered across Slack threads, Jira tickets, model registry metadata, and inference logs. Auditing a model's journey often requires a forensic investigation. Forkit Dev's bet is that by making lineage a first-class, structured citizen from the start, organizations can move from reactive audits to proactive governance, building trust not as an afterthought but as a foundational layer.
Sources
- [Forkit.dev, 2024] AI Passport Infrastructure | https://www.forkit.dev/
- [Forkit.dev, 2024] Forkit Developer Hub | https://www.forkit.dev/developers
- [F6S, 2024] Forkit Dev - F6S | https://www.f6s.com/company/forkit-dev
- [LinkedIn, 2024] Arpita Sarker - Founder & CEO, Forkit Dev | https://de.linkedin.com/in/arpita-sarker
- [LinkedIn, 2024] Syed Amir Hamza - Co-founder & CTO - Forkit Dev | https://de.linkedin.com/in/syed-amir-hamza-a5b5b51b5
- [LinkedIn, 2024] Forkit Dev | LinkedIn | https://www.linkedin.com/company/forkit-dev