DevRay Lab · Founder-led software & AI engineering
Custom AI applications, intelligent automation, web platforms and mobile solutions — designed, engineered and deployed around real operational needs.
Led by Raymond (DevRay) Klanderman · Available for remote collaborations
Founder-led Software & AI Engineering
Common starting points
Most projects start with an operational bottleneck, an AI idea that needs grounding, or a system that has outgrown spreadsheets and manual follow-up.
Services
Each engagement starts with the problem, the people affected and the decision the software needs to support. Technology comes after that.
Custom AI assistants, RAG, document processing and workflow automation.
Production-grade generative AI and agentic systems built with strict guardrails, human-in-the-loop oversight, and rigorous evaluation before deployment.
Evidence: AI Health Chat, Luminae, AutoResearcher, PostShifterPro.
See it live: AI Health ChatRelevant technology
Business portals, dashboards, forms, internal tools and SaaS MVPs.
Robust, scalable web applications and internal tools designed to eliminate operational bottlenecks, with clean data models and role-based security.
Evidence: TutaLearn, Sauti Mtaani Nairobi, Kazi Connect.
See it live: TutaLearnRelevant technology
Flutter cross-platform Android and iOS apps with offline and device capabilities.
Accessible, high-performance Flutter mobile apps engineered from a single codebase, featuring offline-first data sync, native device integrations and structured test coverage.
Evidence: EyeSense accessibility prototype; UrbanSpark AI in development; institutional Pan-African Parliament Flutter/Dart experience (no official endorsement implied).
Read the EyeSense case studyRelevant technology
Cleaning, dashboards, forecasting, recommendation prototypes and model evaluation.
Practical data engineering and predictive machine learning models with honest performance bounds, designed to guide concrete business decisions.
Evidence: Movie Recommendation Project: about 10 million ratings, public-leaderboard RMSE 1.21 → 0.82 across eight submissions. Public-leaderboard context only; not a customer KPI.
Read the recommender case studyRelevant technology
Additional capabilities
Multimedia live streaming and video production: live radio and TV streaming, scheduled programming, SEO-friendly publishing, RSS news feeds and PWA support.
Clear technical content — tutorials, documentation and demos — that helps teams adopt and extend complex systems with confidence.
Custom WordPress solutions: theme customisation, page builders, SEO optimisation, plugin integration, WooCommerce, speed and security hardening, and maintenance.
Social media strategy, content creation and curation, community management, analytics and reporting, brand awareness, and paid advertising management.
Selected work
A selection of live systems across health, education, accessibility, civic technology and machine learning. Each case study separates what was built from what has actually been verified.
PLP Hackathon WinnerPublic beta · Independent product
Problem: Multilingual health information is hard to access reliably, especially on low-bandwidth connections and messaging platforms.
Built: A live multilingual AI health-information assistant on the web and Meta WhatsApp Cloud API, offering educational triage support, multimodal image understanding, vital-sign tracking and physician-ready consultation summaries.
Live · active development · Independent product
Problem: Kenyan learners need curriculum-aligned lessons, assessments and affordable access that works on mobile and low-cost connectivity.
Built: A KICD-aligned CBC platform with strand lessons, term assessments, an AI tutor, WhatsApp reminders and quizzes, and bilingual English/Kiswahili narration.
Live hackathon build · Mistral hackathon project
Problem: Researchers spend hours manually critiquing methodology, auditing datasets and drafting grant proposals from academic papers.
Built: A five-agent research-paper intelligence platform — OCR, vision, critics, synthesis and grant — that critiques methodology, audits datasets, proposes experiments and drafts NSF/NIH-style grant outlines using Mistral models.
Live web app · Civic platform for the Office of the Nairobi Woman Representative
Problem: Residents need a clear way to submit structured feedback, and a constituency office needs aggregated insight without automating decisions that require human judgement.
Built: A bilingual resident-feedback platform that aggregates submissions and uses AI to suggest themes, with every AI suggestion reviewed by staff. PII is scrubbed before any AI call, and there are no automated eligibility decisions.
Mobile prototype · Innovation showcase project
Problem: Blind and low-vision users need practical visual assistance that respects privacy and works when connectivity is limited.
Built: A Flutter Android prototype with offline OCR, speech and vision assistance, a Cloudflare Worker backend, and a TalkBack-first design that asks for consent before any image leaves the phone.
Competition entry · Programme final project (optional)
Problem: Viewers need personalised movie recommendations, and public-leaderboard scores do not translate directly to real-world KPIs.
Built: A hybrid content-based and collaborative-filtering recommender built on about 10 million ratings, using SVD matrix factorisation and LightGBM. Public-leaderboard RMSE improved from 1.21 to 0.82 across eight submissions.
Why work with DevRay
DevRay is founder-led. That means the person who scopes the work is the person who builds it, and the trade-offs are explained in plain language rather than hidden in a handoff.
You work directly with the engineer doing the work. Discovery, architecture, implementation and handover stay in one accountable thread.
Product engineering and applied AI are designed together, so the model, data and interface are not bolted together at the end.
Consent, data minimisation, human review and accessible interaction patterns are considered part of the product, not a later patch.
Tools are chosen for the job and the maintenance budget. Technology names are evidence of work, not a shopping list.
Testing, evaluation and monitoring are scoped to the impact of failure. Claims are qualified when they are not independently verified.
Decision points, deliverables and change control are written down, with iterative previews and documentation at handover.
What to expect
Commitments are qualified, not inflated. Scope, assumptions and change control are written down so both sides know what is being delivered.
How we work
Every engagement follows the same decision points, adapted to the size of the work. You should always know what is being built, what is being tested, and what happens next.
Understand users, constraints, the business decision behind the build, and what success actually looks like.
Typical deliverablesDiscovery notes, open questions and a recommended scope.
Turn discovery into a concrete scope with deliverables, assumptions, risks and a written proposal.
Typical deliverablesScoped proposal, milestones and change-control approach.
Validate the riskiest workflow early — the model behaviour, the core interaction or the integration.
Typical deliverablesWorking prototype or technical spike with findings.
Implement the product in reviewable increments, with previews and decisions documented as the work progresses.
Typical deliverablesIterative builds, preview links and implementation notes.
Test the workflows that matter, check accessibility and security basics, and prepare deployment.
Typical deliverablesTest results, deployment checklist and launch build.
After launch, agree what maintenance or iteration is useful rather than assuming an ongoing retainer.
Typical deliverablesHandover documentation and an agreed support path.
Timelines and costs are quoted after discovery. No guaranteed delivery date or support SLA is implied before the scope and assumptions are agreed.
About the founder
DevRay is the engineering services practice of Raymond (DevRay) Klanderman. Every project is handled directly with senior engineering ownership — spanning architecture, data models, interface design, safety boundaries and reliable deployment.
Software & AI Engineer · Nairobi, Kenya — available for remote collaborations
On client engagements and independent systems, I work across the full development stack and the applied AI lifecycle: data schemas, APIs, interfaces, automated testing and cloud infrastructure, paired with prompt design, agent workflows, model evaluation and privacy guardrails.
Institutional experience
Media Specialist (contract)
Contract work included building a Flutter/Dart cross-platform Android and iOS application, integrating livestreamed Pan-African Parliament sessions, and structuring official Pan-African Parliament content for mobile — working with delegates and multidisciplinary teams. This is described as institutional experience only; it does not imply official Pan-African Parliament product endorsement, production rollout or authorization.
Kubernetes and Cloud Native Associate (KCNA)
Linux Foundation
Verify credential ↗ (opens in a new tab)Foundation Certificate in Generative AI
BCS, The Chartered Institute for IT
Verify credential ↗ (opens in a new tab)Questions
If your question is not covered here, send it through the project form. I will answer directly rather than route you through a sales sequence.
Ask a questionAI applications and automation, web platforms and internal tools, Flutter mobile apps, and data analytics or applied ML work. Smaller supporting capabilities include technical documentation, WordPress, livestreaming and digital media. If a request is outside that range, I will say so and suggest a better fit where I can.
Timelines are quoted after discovery, because they depend on scope, integrations, data access and review cycles. A focused prototype is a different commitment from a production platform. I do not publish guaranteed timelines before understanding the work.
Pricing is agreed from a written scope after discovery. Smaller prototypes and larger builds are structured differently, and assumptions are stated in the proposal. There are no hidden price ranges on this site because a credible number depends on the actual work.
Yes. DevRay is based in Nairobi, Kenya, and works remotely with clients and collaborators. Calls can be scheduled, and most collaboration happens asynchronously through shared previews, written updates and review points.
Ownership and licensing are agreed in the project contract. Client deliverables and any reusable internal tooling are treated separately, and private source code is never published. The exact position is written down before work starts.
Often, yes — after a review of the architecture, dependencies, tests and deployment path. I will tell you if the existing system makes the requested change risky or uneconomic, rather than pretending a rewrite is always the answer.
The aim is to minimise data sent to model providers, avoid unnecessary personal data, use consent where required, and keep a human review step for consequential decisions. Provider data policies still apply. I do not claim formal security certification unless one has actually been obtained.
Testing and maintenance can be scoped with the build. The depth should match the risk: critical flows need more evidence than a marketing page. Ongoing support is agreed explicitly; there is no implied 24/7 support or service-level guarantee.
I will say so early. Where possible I will point you toward a more suitable approach or type of specialist. A short discovery conversation should leave you clearer about the next step even if we do not work together.
Start a conversation
A short, specific description is more useful than a perfect brief. Share the problem, the people affected, any deadline and what you have already tried.
Nairobi, Kenya
Available for remote collaborations. Calls can be scheduled.
I will review your inquiry and follow up by email to discuss next steps. No guaranteed response time or support SLA is implied.
Your details are used to respond to this inquiry and are delivered through Formspree. Read the privacy notice. Do not send passwords, API keys or sensitive personal data through this form.
