WorkSenior Full-Stack & AI Specialist (Contract)

Iluy.ai

A controlled AI assistant that lets users set standards once and apply them across every answer.

The challenge

Iluy’s public product problem is that general AI tools are tuned for broad defaults, not for the standards of a specific family, school, organization, or professional workflow. The product needed to make those standards configurable, persistent, and clear.

The engineering challenge was to combine top-model access, content filtering, custom prompt profiles, model preferences, accuracy context, and SMS/WhatsApp access without turning the product into a pile of disconnected AI settings.

Role: Senior Full-Stack & AI Specialist
Platform: Controlled AI assistant SaaS
Stack: Next.js, NestJS, LangChain, Vercel AI SDK, LLM routing

Product impact

1,500early users highlighted publicly

Webplus SMS and WhatsApp access

ModelsOpenAI, Anthropic, Google, Grok, and more

AI product architectureFull-stack implementationModel routing and orchestrationModeration and standards workflowsPrompt profile and settings UXSMS and WhatsApp access

Next.jsNestJSLangChainVercel AI SDKLLM routingContent moderation

Iluy standards setup screen — Standards can be configured in plain language and applied to the assistant experience.

Iluy content moderation settings — Moderation settings give users control over filtered categories and custom rules.

Product foundation

AI that respects user standards

Iluy’s current public positioning is built around control: users set a standard, Iluy applies it across answers, and the assistant keeps unwanted content out of the experience.

The product work had to make that control visible without making the chat feel heavy. Standards, moderation categories, custom rules, and locking flows all needed to feel like part of one assistant rather than a separate compliance console.

Iluy standards setup modal — The standards setup flow turns content rules into a clear product decision.

Delivery

Many models, one assistant experience

Iluy runs on leading AI models and lets users auto-select or choose models directly. The product surface needed model power without asking users to juggle separate subscriptions, model switches, or provider-specific workflows.

The implementation connected the chat experience, model routing, response modes, settings, and feedback loops so model choice could stay useful without overwhelming the core conversation.

Iluy chat interface with answer controls and model context — The chat surface keeps multi-model behavior inside a familiar assistant flow.

Channels

Access beyond the browser

The public product supports use on the web and positions SMS and WhatsApp as access paths for people who want AI available without opening a full app every time.

Those channels add product complexity: accounts, preferences, moderation standards, and conversation context need to remain consistent even when the user moves between interfaces.

Iluy SMS and WhatsApp settings — SMS and WhatsApp settings extend the assistant beyond the browser while keeping one controlled account model.

Result

Outcome

A controlled AI assistant foundation with standards, moderation, multi-model access, custom profiles, and web, SMS, and WhatsApp usage paths aligned around one product promise: AI that answers to the user’s standards.

Next projectAorForms.comLed a compliance-focused e-signature SaaS from early build to 20,000+ users across document workflows, templates, SMS/email sending, and agent operations.