Skip to content
Joshua Heller
DE EN

I take your AI from prototype to production.

I'm Joshua — AI Engineer, founder of TAISC, and working exclusively in AI for 3 years. As a Forward Deployed Engineer I build directly in your codebase. As Fractional CAIO I lead your AI strategy. No slide decks. Real systems.

Joshua Heller, AI Engineer and Fractional CAIO, at his laptop

Trusted by teams at

BCT Technology Logo
Bitwerft Logo
Onventis Logo
WM aquatec Logo
STIHL Logo
i40 Logo
S-IT Logo
mimacom Logo
ALL:AIRT Logo
squareNeo Logo
koenic Logo
digitalsprung Logo

Who I am

Who I am and what I do

I hold a B.Sc. in Digital Enterprise Management, started as a developer at Daimler Trucks, was AI Engineer at generic.de, built my own SaaS in 2023 (imagetocaption.ai, 350,000+ users, solo, no funding), and now run TAISC.

I know both sides: enterprise software with compliance, legacy, and 10-year roadmaps — and solo SaaS where every day counts. That mix helps me quickly see where AI actually makes a difference in your project.

My mission: making mid-market companies, enterprises, and software firms more competitive with AI. Not through buzzwords. Through code that runs.

More about me
Joshua Heller speaking at an AI event

My service areas

Build AI understanding

AI Workshop, AI strategy, and use-case discovery. You know AI matters — but don't know where to start.

Learn about AI Workshop →

Build AI systems

Forward Deployed Engineer and MVP Service. You know what you want to build — you just need someone who actually delivers.

Learn about FDE →

Lead AI strategy

Fractional CAIO and AI Sparring. You need part-time AI leadership without spending $200k+ a year on a full-time Chief AI Officer.

Learn about Fractional CAIO →

"Most AI projects don't fail because of the technology. They fail because of the distance between the people writing the strategy and the people writing the code. I close that gap — because I do both."

— Joshua Heller

350k+
Users on own SaaS products
5+
Years as Software Engineer
3
Years exclusively in AI
1M+
LinkedIn Impressions / Year

Does this sound familiar?

Most teams I talk to face similar challenges. Not because they work poorly — but because AI is a new discipline where the rules are still being written.

Prototype works, product doesn't

You built a notebook, a ChatGPT integration, a RAG demo. It runs locally. But getting it to a live system is a mountain your team keeps getting stuck on.

Your team is strong, but not deep in AI

Frontend, backend, DevOps — all covered. But vector DBs, LangGraph, eval pipelines? That's where practice is missing. And hiring takes 6 months.

Architecture is a gut-feeling call

"Should we use OpenAI or Anthropic?" "Do we need an agent or will a prompt chain do?" Nobody on the team can answer that from experience — so you're rolling the dice.

Your consultants deliver slide decks

You got a 100-page strategy document. Reading it takes a week. Implementing it takes 12 months. Lines of code written: 0.

Leadership is waiting for results

"Where's our AI ROI?" — great question. But who on the tech team has the bandwidth to figure that out on the side? And who translates it for the C-suite?

A full-time CIO isn't in the budget

You need someone to own the AI strategy. But $200k+ a year for an experienced Chief AI Officer isn't justifiable — and it would take 9 months to find one.

How can I help?

I build inside your codebase

Weekly sprints or retainer. I become part of your team, build concrete AI features to production, document everything, and step back once your team can run independently.

from €200/h Learn more →

My principles

01

I tell you how it is.

If your use case doesn't need AI, I'll say so. If your tech stack plan is a mistake, I'll say that too. No buzzword bingo, no sugarcoated roadmaps.

Honest assessment, even when it's uncomfortable.

02

I deliver code, not decks.

Every engagement ends with an artifact: a working MVP, an architecture doc, a piece of production code, a roadmap with tickets.

Impact over slides.

03

I make myself obsolete.

My goal isn't to embed myself in your team. It's to enable you so thoroughly that you can move forward on your own. Knowledge transfer is part of every project.

No lock-in, no knowledge hoarding.

04

I don't take kickbacks.

I recommend OpenAI, Anthropic, Azure, AWS, Lovable, n8n, Supabase — based on fit, not commissions.

Objective evaluation, no sales pitches.

How we work together

01

Intro call

30–45 min video call. You explain what's stuck. I listen and ask questions. By the end, we both know if and how I can help.

⏱ 30–45 min · Clarity on fit

02

Discovery

I look at your setup: codebase, architecture, team, existing AI approaches. This produces an assessment — not a 100-page doc.

⏱ 1–2 weeks · Solid understanding

03

Proposal

You get a concrete proposal: hour bank for FDE, fixed price for MVP, retainer for Fractional CAIO. With scope, price, and delivery dates.

⏱ a few days · Written proposal

04

Execution

We start. Weekly calls, ongoing availability, code in your repo. I work with the tools you use — and I'm part of your team for the duration.

⏱ depends on engagement · Code, MVP, Roadmap

How I work

Software companies need someone who asks the right questions.

"The best AI is the one your team understands and can operate."
— Joshua Heller

"Should I use Claude Opus 4.7, an OpenAI model, or Gemini?" "Do I actually need RAG, or will the context window handle it?" "Local LLM — yes or no?" These are the questions I help with.

My daily stack: Claude Code and Claude Cowork, CLI-first, LangChain and LangGraph for agentic workflows — deployable on Azure, AWS, Supabase, or Google Cloud. I work with whatever fits your setup.

I don't take commissions. If I recommend Anthropic, it's because it fits. If I say you don't need AI for a use case, I mean it.

What sets me apart from an agency

An agency lives on billable hours — many clients, many tickets. I work intensively with a small number of clients. That's a different model — and it fits AI better than the agency model. My goal is to make myself redundant.

Frequently Asked Questions

What does working together cost?

It depends on the model. Forward Deployed Engineer from €200/h net (from a minimum commitment, with discount as a retainer). AI Sparring for individual hours from €300/h. MVP fixed price from €8,000. Fractional CAIO as a retainer from €3,000/month. We'll nail down the exact terms in the intro call based on scope.

What industries and company sizes have you worked with?

Across the board: automotive and industrial (Daimler Trucks context, STIHL, WM aquatec), B2B SaaS and software (Onventis, Bitwerft, BCT, mimacom), education (i40), and family office / private equity environments. From solo startups to mid-market companies, scaleups, and large enterprises. Working with multiple clients in the same industry is an advantage — I bring patterns from comparable projects.

Are you an agency?

No. I'm CEO of TAISC – The AI Software Company GmbH. TAISC has a small team of experienced engineers, but my personal FDE and Fractional CAIO engagements are done by me directly — not delegated.

How quickly can you start?

Typically 1–3 weeks after signing. For urgent needs, I often find a gap sooner.

Do you work remotely or on-site?

Primarily remote. For workshops, discovery phases, and critical project phases I can come on-site — based in Karlsruhe/Pforzheim, active throughout the DACH region and available internationally.

What if I don't know what I need yet?

That's exactly what the intro call is for. 30 minutes. No commitment. We figure out what makes sense. Sometimes it's a 1-day workshop, sometimes an MVP, sometimes nothing — and I'll refer you to someone who's a better fit.

Ready to take your AI to production?

30 minutes. You explain what's going on. I tell you if and how I can help. No commitment.

Prefer to write first? joshuaheller@theaisoftwarecompany.com