Artificial General Intelligence
Agentic and multimodal systems that reason, use tools, operate computers, coordinate other agents, and execute work over long time horizons. Applied in Senon AI and VERONICA.
About
Founder and lead engineer at Senon Solutions.
I'm Seniya Mendis. I build AI systems, web applications, developer tools, and native software, and I provide the infrastructure they run on. I work directly with the clients I build for, from Galle, Sri Lanka.
Background
Education and experience.
I started programming in 2017 and founded Senon Solutions while completing my degree.
- 01
Founder & Lead Engineer
Senon SolutionsI founded Senon Solutions in 2024 to build products, models, frameworks, and infrastructure as one connected effort.
- 02
Independent Researcher
AI & systems researchCPU and GPU inference optimisation, custom model-execution instructions, multimodal systems, autonomous agents, and cybersecurity for applied machine learning.
- 03
BSc (Hons) Data Science
University of Plymouth, UKSecond-Class Honours, Upper Division.
- 04
Richmond College
Galle, Sri LankaPrimary and secondary education.
Focus
Areas of focus.
Engineering work at Senon Solutions falls across four areas.
Distributed systems
Global traffic routing, edge compute, failover, and capacity management. Applied in Senon Cloud and M-Core Autoscaler.
Cybersecurity for applied machine learning
Isolating confidential material from third-party models, verifying application integrity during execution, and handling privacy and failure at the architecture level. Applied in VERONICA and M-Core Engine.
Low-level programming
Runtimes written in Rust, native compilation, memory behaviour, and performance measurement. Applied in M-Core Engine, SenonUI, and Project O.
Working together
What a project includes.
A project covers the full path. The problem is worked through first, then architecture, build, and delivery. Maintenance and infrastructure are included: Senon Solutions maintains the systems it builds and operates the infrastructure they run on, so there is no separate team to hire and nothing left for you to keep running.
Verification before application
Project O does not apply a generated Rust extension until it has proven behavioural equivalence against the original implementation and measured a real improvement.
Security as architecture
VERONICA rewrites each conversation in full before it leaves, at a measurable cost in latency. Security, privacy, and failure handling are decided during design rather than added afterwards.
Owning the stack
Senon Cloud, M-Core Engine, and the Senon AI models were built in-house because depending on separate external layers meant limited control over any of them.
Expertise
View all expertiseThe full capability map.
The expertise page covers what I build across AI, product engineering, runtimes, and infrastructure, with each area tied to a system.