Software development & AI integration
Software Developer · Lucerne, Central Switzerland
I'm Serhii Ovsiienko, a software developer in Lucerne. I connect existing systems, make company knowledge useful to AI and automate recurring tasks — from the first prototype to day-to-day operation.
Serhii Ovsiienko
Software Developer
01 — For your business
We start with a concrete task and the systems you already use. The result is a solution your team can use in everyday work — in Lucerne, Central Switzerland or remotely.
Information should not need copying by hand between applications. I connect your software through suitable interfaces and make data available for the next step.
An assistant can draw on approved documents, rules and project knowledge for a specific task. Sources and open questions stay visible to your team.
Prepare data, compile reports or sort incoming requests: we choose a manageable workflow and compare time spent and corrections before and after the pilot.
When existing tools fall short, I build the missing application: an internal portal, a dashboard or a web app — including its backend, documentation and operation.
02 — Current work
I build systems that make accumulated experience useful — for AI agents and for robots.
GBRAIN
I am building GBRAIN as a shared foundation for AI agents: persistent project knowledge, access to external MCP tools and orchestration with tasks, handoffs and checks. For software development and pilots in business workflows.
Active development · Pilot projects
Discover GBRAINVBRL
Git for robot experience: store skills as data, transfer them between robots, merge contributions and remove selected records. With traceable sources and interactive demonstrations. Validated in simulation so far; real hardware is the next step.
Research · Demonstrations in simulation
Explore demos & resultsWe can look at your workflow and assess where automation could help.
03 — Working together
Which task takes time? We clarify the workflow, the systems involved and the outcome you need.
We agree on scope, price, data access and required approvals. A limited pilot gives us an outcome we can assess.
A first prototype works with approved examples. Your team reviews results and open questions; we measure time spent and corrections.
If the pilot proves useful, we set up ongoing operation with documentation, clear responsibilities and support as agreed.
04 — Work
Alongside GBRAIN and VBRL, I build products for market analytics, research, e-commerce and payments.
Market analytics
< 1 ms
response time
Explainable market analytics: Rust vector search over 2.5 M price patterns from ten years of history. Every signal ships with its historical evidence.
Desktop app
6.6 M/s
data points
Native macOS app for strategy research. Rust engine with walk-forward and Monte-Carlo validation and 30+ risk metrics.
E-commerce
< 10 ms
per request
Real-time recommendations for online shops: matches a visitor's click path against completed purchases and surfaces the next best step.
Payments
Self-hosted payment and treasury service with a policy engine, API gateway and hardened key handling.
05 — Experience
Rotkreuz, Switzerland
Backend applications and REST APIs in Python and JavaScript, ETL pipelines for financial and market data, dashboards, model infrastructure integrated with IG Bank and Interactive Brokers, LLM integration.
Remote
Grew AI Price Patterns into a production API service; built Manus Pay.
USA, remote
Reinforcement-learning simulation and optimisation systems; interactive dashboards for monitoring and backtesting.
Ukraine
Built the platform: Rust vector-search core and backtesting engine, data pipelines, web frontend, deployment.
Remote
Quantitative strategies, statistical models and predictive-analytics pipelines in Python and R.
Education: Specialist degree in Finance (Economist, Financial Manager), Interregional Academy of Personnel Management, Kyiv.
06 — About
I started in quantitative teams, where a bug in the data costs money immediately. That shaped how I work: measure first, then build, and make every decision traceable.
I live and work in Central Switzerland, and I value clear agreements, reliability and software that still makes sense in five years.
07 — Contact
Which recurring task takes up your team's time? Tell me about the workflow and the systems you use. Together we can assess where software or AI could help. I am also open to software roles, technical partnerships and conversations about my projects. I usually reply within one business day.