Software Engineer & AI Researcher · St. Louis / Remote

KimathiNjoki

Transforming ideas into intelligent solutions, building at the intersection of software engineering and artificial intelligence.

Now · Full-stack & applied AIField · DSS / RAG / GenAI--:-- CTScroll ↓
01 · About

Building with intention

Engineer first, researcher always

I'm a full-stack software developer and AI researcher specialising in decision support systems1 and collective decision-making models. I love building APIs, desktop and server software, with a focus on creating innovative solutions while fostering collaborative environments.

I'm passionate about AI and its integration into preexisting systems, particularly generative AI, model distillation2, and consensus-building technologies.

1Systems that help groups decide, not just dashboards that watch them.

2Big-model judgement, small-model budgets.

Kimathi Njoki

Software Engineer & AI Researcher

  • Python
  • Ruby
  • JavaScript
  • TypeScript
  • Java
  • Kotlin
  • HTML5
  • CSS3
  • React.js
  • Tailwind CSS
  • GraphQL
  • Node.js
  • Express.js
  • Ruby on Rails
  • Django
  • Flask
  • .NET
  • PostgreSQL
  • MySQL
  • MongoDB
  • ChromaDB
  • AWS
  • Google Cloud
  • Excel & R
  • Tableau
  • DevOps & CI/CD
  • Git
  • Langchain & RAG
  • Azure
Specimen · K. Njoki+ rev 2026.06
Based
St. Louis, USA · working worldwide, remote-friendly.
Focus
Full-stack development & applied AI: DSS and collective decision-making.
Approach
Project management & agile · AI integration & RAG systems.
Interests
Systems architecture & integrations.
02 · Process

Database to pixel

One engineer, every layer
  1. 01

    Database to pixel.

    One engineer across the whole stack, so nothing gets lost in a handoff. The schema and the button were designed in the same head.

  2. 02

    No black boxes.

    Every layer is code I wrote and can explain, from the migration to the prompt. Typed, lean, readable — fast because there's nothing extra in it.

  3. 03

    AI where it earns its place.

    Models get measured against a baseline and wired with a fallback. If a plain rule does the job, the rule wins.

  4. 04

    Built to be inherited.

    Tested, documented, boring in the right places. The next engineer — or me, two years on — can pick it up cold.

// the stack, assembled · scroll to build

  1. 01Rulenever miss a severe one, never cry wolf
  2. 02Boundaryno source code, no write access
  3. 03Inventoryevery layer, down to the OS
  4. 04MatchOSV, KEV and EPSS, on a schedule
  5. 05Rankexploited now outranks a CVSS 10
  6. 06Gatethe build fails on what's exploited
  7. 07Prove10 real incidents, replayed
  8. 08Shipone binary, CLI and server agree
  9. 09Decideopen until a person decides
03 · Projects

Selected works

5 builds
  • 01 · FamilyUreriThe whole family, in one place.FamilyEncrypted
  • 02 · WellnessKinnectEvery body, every story, one community.CommunityPayments
  • 03 · ConciergeAtelierKenya, arranged around you.LuxuryMembers
  • 04 · SecurityWatchdepHear about the dependency that matters, not the other four hundred.CLICI
  • 05 · AI AssistantJPROA conversational assistant that knows the project, not just the prompt.LLMRAG

Case studies & repos shared on request → happy to walk the code

04 · Services

What I build

Full stack, integrations, architecture
  1. 01

    Full-stack products

    From the schema to the last pixel, with one person accountable for all of it.

    Next.jsPostgresAuthDeploy
  2. 02

    Integrations & APIs

    The tools you already pay for, talking to each other — so data moves without anyone copy-pasting it.

    RESTWebhooksOAuthQueues
  3. 03

    Systems design & architecture

    The diagram before the code: boundaries, data flow, failure modes, and what it costs to run.

    ArchitectureADRsScalingReview
  4. 04

    AI & LLM features

    Retrieval, agents and evaluations wired into real workflows — and measured against a baseline.

    RAGAgentsEvalsGuardrails
  5. 05

    Data pipelines & dashboards

    Ingest, clean, score, show. Pipelines that run on schedule and dashboards people actually open.

    ETLSchedulingPostgresCharts
  6. 06

    Technical advisory

    A senior second opinion on a codebase, a plan or a hire, with the findings in writing.

    Code auditRoadmapDue diligence

// next step

Tell me about your project.

A product, an integration, or the architecture underneath them. A short brief gets you an honest read on approach, effort and risk — in writing, before anything is signed.

05 · Lab

Things I tinker with

Live · rendered right here

Cut and Paste

One ball, one blade, doubling every turn. Tap to burst it.

Flow Field

Particles riding a noise field.

Matrix Rain

The classic, in dashboard green.

More on GitHub ↗

06 · Contact · St. Louis, CT (UTC-5)

Let's buildsomething.

Full-stack builds, integrations, systems design and architecture. Tell me what you're building — you'll get an honest read on approach, effort and risk, in writing.

dfkimathinjoki@gmail.com
Start a project
What are we building? · pick any
Where is it now?
Timeline?

brief tap what fits, then tell me the rest below

© 2026 Kimathi Njoki38.6270° N · 90.1994° W--:-- CTBuilt at the intersection