About GraceSoft
GraceSoft is an AI engineering studio building production-grade AI bots and automation — from conversational agents on WhatsApp and Telegram to retrieval-augmented systems grounded in your own data.
Meet the Founder
GraceSoft is founded by Davina Leong, an AI Engineer and full-stack developer with 10+ years of software delivery experience. Her work spans LLM integration, RAG pipelines, agentic workflows, and AI-powered SaaS — including production AI systems deployed in regulated, high-compliance environments.
Before founding GraceSoft, Davina built a production RAG pipeline and GPT-based chatbots into STARLIMS, a laboratory information management system used by PUB, Singapore's national water agency — surfacing regulatory documentation and workflow guidance for lab technicians in real time, and prototyping an MCP (Model Context Protocol) architecture for AI-database orchestration.
That same rigor shapes every GraceSoft product: Sentinel Concierge integrates WhatsApp Business and Telegram Bot APIs with an LLM conversation layer and business-hours- and public-holiday-aware scheduling logic. Sentinel Cook generates recipes from a single photo using a vision-language model, with a Google Drive integration that turns a user's own recipe library into a RAG-grounded knowledge base. Comfortable across the full stack — from model orchestration and API design to front-end implementation and cloud deployment on AWS.
Experience & Background
- 10+ years of software delivery experience across AI engineering and full-stack development
- Production RAG pipeline + GPT-based chatbots built for STARLIMS, a lab information system used by PUB (Singapore's national water agency)
- Prototyped an MCP (Model Context Protocol) architecture for AI-database orchestration
- Founded and architected GraceSoft's AI product suite — LLM integration, agentic workflows, and automation pipelines deployed to production for SME and enterprise clients
- Business Data Analyst training with BCG Rise 2.0 — graduated top of cohort in SQL, Python, and Data Analytics, with distinction
- Interviewed by BCG to provide input for an AI-powered career tool they were building
- Applied AI for Developers training (Vertical Institute) — advanced prompting techniques that raised AI code-generation accuracy by 40%
- Certified in UX Design (Vertical Institute); cybersecurity training including a PDPA/GDPR-aligned incident response capstone
- Award-winning project (SPINNOVEX Gold)
Why GraceSoft
Most AI pilots never make it to production — they work in a demo, then stall on real users, real data, and real edge cases. GraceSoft exists to close that gap: AI systems built with the same engineering discipline as any other production software, not a prompt bolted onto a chatbot widget.
That standard comes from experience most AI shops don't have: real production AI work in a regulated, high-compliance environment (more on that above), where accuracy and auditability aren't optional. The same rigor shows up in every GraceSoft product, from business-hours-aware booking logic to RAG systems grounded in a client's own documents.
- Production-Grade AI
- AI systems engineered to run in production, not just a demo
- Grounded in Real Data
- RAG pipelines and agentic workflows grounded in your own documents and data
- Compliance-Ready
- Built with the rigor regulated, high-compliance environments demand
Software Naming Convention
A simple naming system so business owners can immediately understand what each product is and what it does:
GraceSoft + Product Name + Function Name (optional)
Think of it like an address — GraceSoft is who built it, the product name is which product line it belongs to, and the function name is what specific job it performs.
- GraceSoft = company name
- Sentinel = product name (umbrella suite for all AI products)
- Concierge = function name (FAQ and appointment booking AI agent)
- GraceSoft = company name
- Sentinel = product name (umbrella suite for all AI products)
- Cook = function name (culinary and menu workflow assistant)