# Raza Rizwan

> The governed record of Raza Rizwan, Forward Deployed Engineer: a portable method for counting how much of any workflow can be automated, what he builds, how he works, and his journey in AI from 2020 to 2026.

- name: raza-sor
- build_id: sha256:d9f989bee8fa4c15cf0f7f1572f9e9a18305ef4e2ce0c3cfb6a5a421ead6dc1d
- dirty: true

## Documents

- [Who Raza Rizwan is](/docs/about/overview): Forward Deployed Engineer who puts a team's software delivery process inside the AI, so 70%+ of a run happens without a person typing instructions.
- [What I build](/docs/about/what-i-build): AI employees, MCP connectors, skills, custom agents, and the AI-native company, explained by concept with software-delivery examples.
- [How I work](/docs/about/how-i-work): One workflow first, success criteria before the build, the server enforces the process, humans at decisions, and every number counted rather than guessed.
- [Now](/docs/about/now): The technology Raza works with today, and why each piece is in the stack. Projects and clients are deliberately absent.
- [Questions people ask](/docs/about/faq): Direct answers on what an FDE does, what a connector is, whether the 70% is real, what happens at a context reset, and why the internals are not published.
- [Contact](/docs/about/contact): How to reach Raza Rizwan, what to bring to a first conversation, and what happens in a workflow audit.
- [The workflow audit](/docs/audit/overview): What happens in the first conversation, what you walk away with, and what it costs you to find out whether automating your work is worth it.
- [What to bring](/docs/audit/what-to-bring): The six answers that turn a vague "can AI help my business" into a step table worth reading, whichever industry you are in.
- [The journey, 2020 to 2026](/docs/journey/overview): How Raza Rizwan went from training image classifiers in notebooks to deploying AI systems that run real business operations, one era at a time.
- [2020 to 2021: predictive AI](/docs/journey/2020-2021-predictive-ai): The first two years, where classic machine learning and deep learning were learned in public, one classifier at a time.
- [2022 to 2023: MLOps](/docs/journey/2022-2023-mlops): Two years spent learning to ship models, with CI/CD, DVC, MLflow, YOLO object detection in production, SageMaker endpoints and the first private client work.
- [December 2023: the generative turn](/docs/journey/2023-the-generative-turn): The week the repositories changed character, from predictive models to generative AI.
- [2024: GenAI APIs and microservices](/docs/journey/2024-genai-apis): The year LLM features were shipped as clean, documented APIs, and the first paid client app was built.
- [2025: AI agents](/docs/journey/2025-ai-agents): The year the job changed from building models to building workers, and the first year of agency work at Canz Marketing.
- [2026: AI employees and MCP connectors](/docs/journey/2026-ai-employees-and-connectors): The Forward Deployed Engineer work, where AI systems hold real jobs inside a business and nine live MCP connectors run an agency's operations.
- [The journey in numbers](/docs/journey/numbers): Every count this record makes about the journey, with the date it was read and the method behind it.
- [Repositories by year](/docs/journey/repositories-by-year): The complete list of public repositories owned by rrizwan98, year by year, with private repositories counted but not named.
- [The method](/docs/method/overview): A portable way to work out how much of any workflow can be automated, which parts must stay with a person, and where to start.
- [Count your own workflow](/docs/method/count-your-own-workflow): How much of my own work can be automated: the four-step procedure that turns any run of work, in any industry, into a tagged step table and an honest percentage.
- [What automates and what does not](/docs/method/what-automates): Domain-free rules for deciding whether a step belongs to the AI or to a person, and the four conditions a step must meet to be automated safely.
- [Worked counts](/docs/method/worked-counts): The same counting method applied to five different businesses, each with its own step table, its own number and its own human gates.
- [From one task to an AI-native company](/docs/method/from-one-task-to-a-company): The ladder from a single automated workflow to a company where work enters by itself and people sit only at the review points.