# Nnenna Ndukwe Software engineer, applied AI researcher, AI ecosystem architect, and Developer Relations Lead at [Qodo](https://www.qodo.ai/authors/nnenna-ndukwe/). I build tools and explain how developers can check the software they create with AI. My work focuses on code quality, engineering standards, and the conditions an agent must satisfy before its changes can move forward. I turn the questions I encounter in engineering into code, experiments, technical writing, and teaching material that other developers can inspect and use. ## Engineering background My background spans software engineering in medical technology, financial technology, and media technology, followed by developer advocacy and technical education. I worked as a software engineer at O’Reilly Media and later in developer relations at Slim.AI and Unleash before joining Qodo. At O’Reilly, I worked with fellow engineers on moving live-event functionality from a monolithic application into services, including registration, event metadata, and the front end. That experience involved keeping an existing product operating while changing its architecture. I discussed the engineering work and my transition into developer advocacy in [Developer Journey episode 264](https://devjourney.info/Guests/264-NnennaNdukwe.html). My work at Unleash connected developer education with feature management and controlled releases. In [Full Stack DevRel](https://cfe.dev/talkshows/december2023-nnenna-ndukwe/), I discussed how my full-stack background, writing, public learning, and community work came together in developer advocacy. At Qodo, my work spans developer education, enterprise enablement, workshops, technical demonstrations, and writing for AI-assisted software development. I connect patterns from practitioner conversations with product feedback and the guidance teams need to evaluate and adopt these tools. My [published work](https://www.qodo.ai/authors/nnenna-ndukwe/) examines how engineering standards, codebase context, independent review, and verification fit into a developer’s daily work. ## Open source projects My public projects approach different parts of the same engineering problem: making intent, authority, and evidence explicit when software is built with agents. - [PolicyNIM](https://github.com/nnennandukwe/policyNIM) provides a policy-aware engineering preflight layer for AI coding agents. It brings policy context into engineering work before a change proceeds. - [Governed Agent Autonomy Patterns](https://github.com/nnennandukwe/governed-agent-autonomy-patterns) explores planning, permissions, tool trust, and independent verification within a bounded agent run. Its scope is the controls around an agent’s permitted actions. - [ThreadLoop](https://github.com/nnennandukwe/threadloop) models an AI-assisted software-delivery task as a governed lifecycle. It preserves task context and connects transitions to current evidence, review, repair, and human decisions. - [Controlled Release Lab](https://github.com/nnennandukwe/controlled-release-lab) is a runnable teaching project for attested builds, hosted deployment, staged feature exposure, and recovery. Its synthetic product catalog makes release behavior observable through request, deployment, and source-identity records. The repository separates completed hosted exercises from workshop rehearsals still pending. - [Software Standards Bootstrap](https://github.com/nnennandukwe/software-standards-bootstrap) combines a CLI with an agent skill to propose repository-specific rules, verification recipes, skills, and an AGENTS.md entry point. Proposals are grounded in repository evidence and remain subject to developer review before adoption. - [redteaming-ai](https://github.com/nnennandukwe/redteaming-ai) is a small, alpha-stage environment for exploring security failures in LLM-integrated applications. It includes a deliberately vulnerable sample application, reproducible local assessment flows, and reporting. It is intended for experimentation and education, not presented as a production-grade security assessment platform. ## Research and evaluation I run experiments on the checks used in AI-assisted development and publish the methods and limits alongside the results. My [TypeScript logic verification experiment](https://nnennandukwe.com/research/admission-verification/) compares testing and formal checking on a standalone admission-logic model. The findings apply to the tested model and workflow, rather than establishing correctness for an entire application. My research also asks how developers should evaluate AI tools beyond a convincing demonstration: what evidence supports a finding, what context the tool used, where human judgment belongs, and whether the result remains valid after the code changes. I connect these questions to practical workflows rather than treating model output as sufficient evidence on its own. My public resources also include [The Software Integrity Evaluation Scorecard](https://docs.google.com/document/d/1HTNZwC4dy-o2rE_yw8I2vx5g7voRT41m/edit), a framework for evaluating AI-assisted code and developer workflows, and [The Enterprise AI Operating Model for Engineering Organizations](https://docs.google.com/document/d/1P4kMvHMZv2vBgPw3EVWJtu8QWmBv9vKT/edit), a vendor-neutral brief on adopting AI while preserving code quality and maintainability. Both are linked from my [public resources page](https://www.nnennahacks.com/work-with-me). ## Teaching and speaking I teach the [AI Code Review course with DeepLearning.AI](https://www.deeplearning.ai/courses/ai-code-review/). It includes building a context engine with chunking, embeddings, and vector search to explore why a review system needs information beyond a code diff. For a practical example, my [quality-first AI coding workshop](https://www.qodo.ai/blog/intro-quality-coding-workflow/) uses a small FastAPI payments service to add a refund endpoint. The lesson connects requirements, repository rules, behavior tests, deterministic checks, independent review, and repair while preserving authorization, idempotency, and API contracts. Selected technical sessions include: - [How to Build Quality Gates into Agentic Coding Workflows](https://ai.engineer/talks/-AKPRSYccJc-build-quality-gates-into-agentic-coding-workflows), a recorded AI Engineer session on checking agent-generated changes. - [Code quality in the age of AI](https://leaddev.com/technical-direction/code-quality-in-the-age-of-ai), a LeadDev workshop co-hosted with Shlomo Dalezman, released June 16, 2026. We covered code-quality governance and its rollout across engineering teams. - [11 Principles for Evaluating AI Dev Tools](https://www.wearedevelopers.com/world-congress/agenda/sessions/1121641), WeAreDevelopers World Congress, Berlin, July 2026. I presented a framework for evaluating generated code and the tools around it. My [post-event recap](https://www.linkedin.com/posts/nnenna-ndukwe_i-feel-honored-i-got-to-take-the-stage-at-activity-7482422722824806400-coTQ) describes the architecture and tools discussed. My [speaking and teaching portfolio](https://nnennandukwe.com) also includes AI Engineer Miami, QCon AI, and AI Engineer World’s Fair in 2026. Earlier work includes AI Native DevCon, All Things Open, Commit Your Code, DevSecCon, dotAI, and DEF CON’s AppSec Village in 2025, alongside mentoring and educational sessions at Harvard and MIT. Recorded sessions cover [autonomous pipelines](https://youtu.be/xH_UJL4NeL8), [separation of agentic concerns](https://youtu.be/VfUTkoYWsPQ), and [moving from AI chaos toward production readiness](https://youtu.be/VykctF7G--Y). These extend the portfolio beyond code review into the design and operation of agentic systems. I also serve on the [AGNTCon + MCPCon Europe program committee](https://events.linuxfoundation.org/agntcon-mcpcon-europe/program/program-committee/). ## Writing and conversations I write [Nnenna Hacks](https://www.nnennahacks.com), a technical newsletter about AI, emerging technology, software engineering, and developer experience. I also publish technical articles at Qodo and co-host [The Agentic Review](https://www.fame.so/shows/the-agentic-review) with Itamar Friedman, discussing AI-assisted development, code quality, and human oversight. Selected writing: - [AI Coding Needs a Centralized Context Plane and Verification](https://www.qodo.ai/blog/context-plane-and-verification/), April 27, 2026. How engineering standards can reach agents and review systems throughout development. - [Moving from AI Code Review to the Outer SDLC Loop](https://www.qodo.ai/blog/moving-from-ai-code-review-to-the-outer-sdlc-loop/), August 13, 2026. How ThreadLoop and Governed Agent Autonomy Patterns address task state, authority, evidence, and progression beyond an individual review. - [Token Efficiency Is an Architecture Decision](https://www.nnennahacks.com/p/token-efficiency-is-an-architecture-decision), July 24, 2026. How the structure of an AI workflow affects its use of context and tokens. - [How I Design My Codex Harness](https://www.nnennahacks.com/p/how-i-design-my-codex-harness), October 4, 2026. A practical account of the environment and workflow around my coding agent. - [Attestation in Loop and Graph Engineering for Software Factories](https://www.nnennahacks.com/p/attestation-in-software-factories), October 11, 2026. How evidence relates to the specific change it evaluated and the decisions a workflow can make from it. - [Verified AI and Formal Verification Are Changing Agentic Engineering](https://www.nnennahacks.com/p/verified-ai-and-formal-verification-are-changing-agentic-engineering), October 25, 2026. An exploration of formal verification in the emerging agentic-engineering workflow. ## Community and recognition I am an [OpenAI Codex Ambassador](https://developers.openai.com/community/codex-ambassadors/directory) and a [2026 Agentic AI Foundation Ambassador](https://www.credly.com/badges/dcf4938e-4024-4819-a152-58760c314b57). My conference recognition includes [AGNTCon + MCPCon Japan speaker](https://www.credly.com/badges/e81666fb-e335-4fdc-aa41-c1d0919fe91e) and [AGNTCon + MCPCon Europe speaker](https://www.credly.com/badges/3f2fe21b-951e-48d6-9e6b-2ba727e46895) credentials. My [public biography](https://www.nnennahacks.com/about-me) records my 2019 Google Women Techmakers scholarship and the National Society of Black Engineers Boston 2018 Engineer of the Year award. I studied Computer Science at Boston University and am a Resilient Coders alum. My community work has included Girl Develop It and Black Girls Code, discussed in my [CFE.dev conversation](https://cfe.dev/talkshows/december2023-nnenna-ndukwe/). ## Find me [Website](https://nnennandukwe.com) · [GitHub](https://github.com/nnennandukwe) · [LinkedIn](https://www.linkedin.com/in/nnenna-ndukwe/) · [X](https://x.com/nnennahacks) · [Newsletter](https://www.nnennahacks.com) Profile facts checked October 29, 2026.