Building APIs in Seconds: How apichap blends APICHAMP Core and AI for automated API generation – Techlead Story with Michael Weißenböck

Context: Inside a lean engineering organization with an ambitious vision

In our DevJobs.at Techlead Story “Michael Weißenböck, CTO & Co-Founder von APICHAMP” (Speaker: Michael Weißenböck, Company: apichap), we got a clear-eyed look at how a very small team fuses product vision, engineering practice, and hiring expectations. The setup is deliberately lean: two people drive the dev team today—Dominik (CEO & Founder) and Michael (CTO & Co-Founder). The ambition, however, is bold: not just to build APIs faster, but to make them fully automated and available in seconds.

What stood out to us is the combination of focused product development (APICHAMP Core) with a research-backed approach to the new AI component—funded through an approved project by the Austrian Research Promotion Society (Forschungsfördergesellschaft Österreich) and implemented in collaboration with the SCCH Software Competence Center Hagenberg. For ML engineers, this is meaningful: the work goes far beyond simple model consumption and reaches into hosting, fine-tuning, training, and integrating language models into a developer-facing product.

Vision: From desired interface to live API—automated end to end

APICHAMP pursues a sharp mission: radically simplify and accelerate the creation of APIs. The core product exists today—APICHAMP Core—which takes a config file and turns it into a running API. On top of this, apichap is building the AI component to automate the path from a user’s desired interface (“Wunsch-Schnittstelle”) to deployment—through intelligent analysis of the data source, semantic matching with the specification, and a fully generated configuration for the core.

The product idea is deeply developer-friendly and addresses a real bottleneck: APIs are the connective tissue of modern software landscapes, but building them takes time, coordination, and repetitive work. APICHAMP aims to compress that into seconds while preserving quality by deterministically linking the specification with the data structure.

Product architecture: APICHAMP Core meets AI

APICHAMP Core

• Technology: Java and Spring Boot.
• Principle: A config file describes the interface; the core turns that description into a production-ready API.

AI component

• Implemented in Python, coordinated via a middleware.
• Focus: NLP with LLMs and embedding models.
• Infrastructure: Vector databases for storing and retrieving embedded information.
• Responsibilities: LLM hosting, fine-tuning, training for use-case-specific requirements, systematic experimentation with different models.

This division of labor is elegant: the core remains a robust, deterministic engine in Java/Spring Boot, while the AI layer brings semantic understanding and config generation. Together, they create a system that moves from a well-defined spec and a described data structure to a running API—without handcrafting boilerplate.

Research foundation: Approved project with SCCH

apichap anchors its AI effort in an approved project with the Forschungsfördergesellschaft Österreich and the SCCH Software Competence Center Hagenberg. That matters because it opens room for actual problem-oriented research whose results are expected to flow directly into the product.

• Research topics are deliberately tackled.
• Findings are transferred to product features.
• The applied research setting brings clarity about trade-offs—from hosting hurdles to fine-tuning strategy.

For candidates, this means you help shape both research and product—and see real-world impact in short cycles.

From input to live API: The APICHAMP flow

Michael’s description makes the vision concrete. APICHAMP centers on two inputs:

1. API specification (desired interface)

• Users define what the interface should look like and document endpoints and structures.

1. Data source

• Supported sources include databases, other REST APIs, XML files, or CSV files.
• Crucially, APICHAMP needs the data structure, not the raw data.
• For databases: a list of all tables and fields.
• For REST APIs: the API documentation.
• For CSV: the header row to identify fields.

From there, the process is:

• Analyze and interpret fields: What semantic content does each column/property represent?
• Match to the desired interface: Bring the semantically understood fields in line with the spec.
• Generate the config file: Produce the config the APICHAMP Core uses to deploy the API.

The AI component accelerates the “interpretation and matching” step—precisely where traditional workflows involve long meetings and fragile manual mappings.

Engineering culture: Lean, focused, tightly synchronized

apichap works with Scrum, weekly sprints, and very close coordination. In a team this small, that is more than process—it’s a tempo system that enables quick reactions and iterative product steering.

• Weekly cadences keep feedback loops short.
• Close coordination reduces context switching and prioritization overhead.
• Decisions and research findings flow into implementation quickly.

Michael also notes that as apichap grows, the organization will evolve. The cultural core—short paths, direct exchange, focus—comes across as a deliberate strength.

Tooling and environment: Autonomy with solid guardrails

apichap sends a strong signal to engineers: choose your environment.

• Operating system: up to you.
• IDE: up to you (the team often uses IntelliJ).
• Codebase: GitLab.
• Delivery approach: Pipeline-Driven Development, ensuring quality and automation.

This setup enables personal flow (your OS and IDE) while keeping the team’s delivery reliable (pipelines, GitLab, established practices).

AI focus: NLP, LLMs, embeddings—and the hard parts of hosting

The AI work tackles the real-world hurdles:

• Which language models serve which use cases best?
• How do you reliably and securely host LLMs?
• How do you approach fine-tuning and training so models truly grasp the vocabulary and structures of both data sources and specs?
• How do embedding models and vector databases get orchestrated for stable, reproducible semantic retrieval?

Michael characterizes the AI effort as strongly experimental. That’s where value emerges: through structured trials, comparable results, and conversion into product features.

Hiring now: ML engineer bridging research and product

The current hiring focus is a ML Engineer. From the session, it’s clear what makes this role distinctive:

• You’ll work on the AI component of APICHAMP.
• You’ll operate across NLP/LLM/embeddings and bring models into a production architecture.
• You’ll help automate the matching of API specs and data structures—including generating the config file for the core.
• You’ll translate findings from an approved research project into shipping features.
• You’ll handle hosting, fine-tuning, and training of language models for concrete use cases.

For engineers who want neither a purely prototype lab nor a rigid product assembly line, this is a rare blend: research depth plus product responsibility.

Hiring process: Personal, focused, and fast

apichap keeps recruiting deliberately simple:

• Phase 1: Personal conversation to get to know each other and assess the collaborative fit.
• Phase 2: Technical interview to validate skills and role alignment.

Recruiting is done by the team itself—no layers, no prolonged loops. Candidates get fast decisions and direct access to the founders.

Work model: Hybrid, with a preference for in-person exchange

The team works hybrid—with an office in the Tabakfabrik Linz and home office as an option. In-person collaboration is valued for the speed of shared thinking and deciding, while remote flexibility supports focused work.

apichap also offers an employee participation program and “a lot, a lot, a lot of room” to shape both product and organization.

Why apichap? Concrete reasons we took away from the session

• Clear, measurable mission: APIs in seconds—not just a slogan, but technically grounded.
• Mature core: APICHAMP Core exists and deploys APIs from a config file.
• Research strength: Approved project with the Austrian funding agency and collaboration with SCCH.
• Impactful stack: Java/Spring Boot for stability, Python for rapid AI iteration.
• Substantive AI: NLP/LLMs/embeddings, vector databases, hosting, fine-tuning.
• Engineering culture: Scrum, weekly sprints, tight coordination, pipeline-driven delivery.
• Autonomy: Choice of OS and IDE (IntelliJ is common), real ownership over your day-to-day.
• Growth stage: Plenty of influence on product, stack, and organization.
• Simple process: Two interviews, direct contact, fast decision-making.
• Location and setup: Tabakfabrik Linz + home office—with emphasis on in-person knowledge sharing.

Who will thrive here

• ML engineers who want to do more than prompting: host, train, fine-tune, and operationalize models.
• Builder profiles who enjoy end-to-end responsibility—from data models to deploy pipelines.
• Generalists with an AI focus who value the interplay of research and product and are comfortable around a Java/Spring ecosystem.
• Startup-minded engineers who prefer short paths, crisp outcomes, and a reliable team dynamic over big-department structures.

Details that inspire confidence

• The precision of the input description (spec + data structure) and the subsequent steps (interpretation, matching, config generation) shows deep process thinking—this isn’t “AI sprinkled on top,” but a thoughtful blend of determinism (core) and semantics (AI).
• The explicit note that APICHAMP needs structures, not raw data, supports a strong architectural stance: preserve data ownership and reduce integration friction.
• The emphasis on hosting and fine-tuning recognizes the tough parts often glossed over—this is where product quality is won.

What day-to-day success looks like

Even as apichap remains small, the bar is high. If you join, you’ll likely enjoy experimenting, measuring outcomes, and turning results into product behavior quickly. Weekly sprints act as a productivity engine: if you want to ship, you’ll find a fitting structure.

The social component matters especially in small teams—trust is a productive asset.

Looking ahead: Growth with adaptability

Michael is clear: as apichap grows, the organization will adapt—not dogmatically, but pragmatically. That sets up a rare opportunity: as an early team member, you help shape not just code and models, but also rituals, roles, definitions of done, and quality standards. If you value this dynamic, you won’t just be “hired” at apichap—you’ll have impact.

Takeaways from “Michael Weißenböck, CTO & Co-Founder von APICHAMP”

• apichap combines an available core product (APICHAMP Core) with an ambitious AI layer that turns specs and data structures into working APIs.
• The team uses Scrum, weekly sprints, and tight coordination—short paths, fast decisions.
• An approved research project and the SCCH collaboration give the AI work depth and direction.
• The open role is a ML Engineer who does more than evaluate models: you’ll host, fine-tune, and integrate them into production.
• Work model: hybrid, office at Tabakfabrik Linz, home office, employee participation program, flexibility in OS/IDE.
• Tooling: GitLab and pipeline-driven development.

Conclusion: An invitation to builders

“Michael Weißenböck, CTO & Co-Founder von APICHAMP” conveys a crisp product vision and an engineering culture that values substance over show. If you want to truly automate APIs in a compact, highly capable team—rather than building paper prototypes—this is a stage where research and product meet.

apichap offers a rare combination: greenfield opportunity in the AI layer, a robust core product, and meaningful influence—technically, organizationally, culturally. Echoing Michael’s sentiment: when the interplay fits, a lot becomes possible.

• 

  

  Negar Layegh, Machine Learning Engineer bei APICHAP

  Machine Learning Engineer bei APICHAP Negar Layegh gibt im Interview Einblicke in ihren Werdegang bis hin zur aktuellen Arbeit, welche Herausforderungen im Machine Learning es gibt und gibt Tipps für Neueinsteiger.

  Watch now