From zero AI experience to paid freelance AI work

The Person

Alex is a full stack developer in Colombia. Before this cohort he had built software before, but had no job, no project, and no real experience in applied AI.

The Before

He looked for AI courses before finding MasterDexter and could not find anything that covered applied AI end to end.

Most courses mixed in machine learning theory and deep learning fundamentals without ever getting to a real, deployed project.

His words: "I didn't have the basis to know how to explore, where to explore, what to find, what to expect. Very lost in this world."

He found Himanshu through a mentor platform and joined after seeing the curriculum was current with what was actually changing in the field.

"I couldn't find a course that covered everything in applied AI. They mix in machine learning and deep learning theory but they don't launch real production projects. This cohort actually does."

The Hardest Part

The local RAG system, Neural Vault.

Local language models underperformed compared to hosted, larger models, so he had to rework the ranking and evaluation logic to get retrieval accuracy up. He also tried building his own evaluation approach instead of using the LangSmith tooling provided, which cost him extra time and introduced a bug.

He worked through it. His approach: solve it himself first, then check the Notion resource library from the cohort to verify whether his approach was correct or needed to change.

What He Actually Built

• A local RAG pipeline (Neural Vault track), with custom retrieval ranking and evaluation work.
• An AI agent orchestration approach, adapted into a real funding-proposal workflow rather than copied as-is from the cohort artifact.

The Paid Work That Followed

Project 1 (completed and paid): An agent-orchestration system that takes a raw problem description and structures it into a formal funding request, submitted to a Colombian entity that approves and funds public projects.

Project 2 (deployed on AWS): Digitizing clinic medical histories for a healthcare client.

Both projects came directly from skills built during the cohort, before his public portfolio was even finished.

The After

Before this cohort: no AI job, no AI project, no AI experience.

After: 2 real clients. 1 paid. 1 deployed on AWS.

"This cohort doesn't only give you demos. It immerses you in real world AI production systems and teaches you the basis to survive and give your best."

Where He Is Now

Catching up on the remaining Cohort 1 artifacts he watched on recording rather than building live. He was deep in client work during that stretch.

Open to running a speaker session in a future cohort about his real client project experience once it is complete.

Being considered for placement as an AI Engineer through AITalentStudio.

Best for: Developers with no AI background who want to go straight to freelance client work, not just a job application.