2011–2014
Web development practice
Independent projects
Started with WordPress brochure sites and moved into business web development, Bitrix and 1C integrations.
YURI ELISEEV · AI ENGINEER
Give me a hard problem.I’ll build a production-ready product.
I design and build complex digital products from the business problem and architecture to full-stack implementation, AI, data and production operation.
LEAD AI ENGINEER
Yuri Eliseev
AI Systems Architect · Full-Stack Product Engineer
ARCHITECTURE · AI · FULL-STACK · PRODUCTION
15
years in IT
since 2011
PRACTICE → ARCHITECTURE
A long engineering path without losing sight of the business problem
I have worked in IT since 2011. I started with WordPress brochure sites and later worked with Bitrix and 1C. Today I architect and build AI platforms, SaaS products and automation systems. Experience across successive technology generations helps me separate durable product logic from temporary tools and choose a solution that can be developed over time.
2014–2022
Technical Director
ELLUR LLC
Led the technical direction and delivery of websites and commercial web systems built with Bitrix.
2022–2026
Full-Stack Developer
CYBIC
Product architecture and full-stack implementation of AI platforms, SaaS products, data systems and production infrastructure.
END-TO-END OWNERSHIP
I lead a complex product as one engineering system.
Product architecture, interface, backend, AI, data, integrations and production remain connected by one technical logic. Assistants can accelerate research, testing, documentation and repetitive operations, while key decisions, quality control and responsibility for the result remain with me.
Business problem
I clarify the actual constraint, the decision-maker and the measurable outcome before choosing technology.
Architecture
I define system boundaries, the data model, integrations, risks and a realistic delivery sequence.
Implementation
I connect the interface, backend, AI layer and infrastructure into one maintainable product.
Production
I validate, deploy, observe and develop the system against the result the business needs.
WHAT YOU CAN DELEGATE
From architecture to a working product
I can join at the point where the problem is still unclear, or take over an existing product that needs a stronger architecture, an AI layer or a production-ready implementation.
Product and system architecture
I turn an uncertain brief into a clear product structure: boundaries, priorities, data flows, integrations, risks and delivery stages. Architecture can be delivered as a standalone package for another team or as the foundation for my full implementation.
AI systems, agents and RAG
I design controlled AI workflows: model and provider abstraction, tools, memory, context, retrieval, evaluation and human approval. The result is an operational product layer, not an isolated model demo.
Full-stack product development
I build interfaces, APIs, business logic, realtime workflows and external integrations as one delivery. Product decisions remain connected to implementation, which reduces coordination loss and shortens the path to a working release.
Data, infrastructure and operation
I own data models, migrations, permissions, deployment, monitoring and post-release iteration. A product is complete when it is secure, observable and maintainable in daily operation.
RESEARCH AND APPLIED SCIENCE
Engineering work in medicine and industry
My engineering practice includes university research, a medical conference presentation and an intellectual-property database for the timber industry. Each item below links to the relevant institution or official record.
2024–2025
Technical Lead in a university research group
Saint Petersburg State Pediatric Medical University · Department of Obstetrics and Gynecology
I led the technical direction of research at the intersection of artificial intelligence, clinical data analysis, obstetrics and gynecology.
28–29 March 2024
Speaker at a medical research and practice conference
City Consultative and Diagnostic Center No. 1 · 2nd City Scientific and Practical Conference with International Participation · Saint Petersburg
I presented “Applying artificial intelligence to the analysis of pregnancy and childbirth histories in women with gestational diabetes mellitus.”
6 July 2026
International Classifier of Commercial Timber
Applied research for the timber industry · state-registered database
I developed the database as an intellectual property asset. Its state registration is recorded in the official FIPS registry under the name “Yuri A. Eliseev’s International Classifier of Commercial Timber.”
EDUCATION · MOSCOW
University of Artificial Intelligence
2024 · 1.5-year programme. Specialisation: artificial intelligence, neural networks, Data Science and machine learning.
ENGINEERING STACK
Technologies behind full-cycle delivery
The stack is organised by engineering area. Tools are selected for the product and connected within one architecture rather than treated as isolated keywords.
AI / ML
PythonAnthropic APIOpenAI APIYandexGPTGigaChatLocal LLMLangChainLangGraphRAGMCPAI AgentsEvaluation
AI products, agents and controlled model workflows
FRONTEND
TypeScriptReactNext.jsViteTailwind CSSTanStack QueryZustandReact FlowPWA
Product interfaces, dashboards and control systems
BACKEND
PythonFastAPIDjangoNode.jsNestJSRESTGraphQLWebSocketPostgreSQLRedisSQLAlchemyRBAC
SaaS logic, integrations and realtime operation
DATA SCIENCE
NumPyPandasSciPyscikit-learnJupyterSQLETLFeature EngineeringVector DBClassification
Data pipelines, analytical models and ML workflows
INFRA / DEVOPS
LinuxDockerDocker ComposenginxVPSGitHub ActionsCI/CDSSLsystemdS3Monitoring
Deployment, observability and production operation
WEB3 / BLOCKCHAIN
SolidityEVMOpenZeppelinHardhatFoundryethers.jsviemWalletConnectIPFSSmart Contracts
Architecture and decentralised services
SELECTED PRODUCTS
Products tested by real users and operation
Three projects show the range of responsibility: AI SaaS, a local-first Android product and an HRTech data platform.
AvtoUM
Working AI SaaSDesigned and built end-to-end
AI SaaS for customer communication: dialogue, qualification, lead capture, CRM and operational control.
- AI employee, business knowledge and need qualification
- SaaS, CRM, billing and administration
- realtime Pulse, human takeover and AI cost control
SvoyUM
Released Android productProduct built end-to-end
A local-first system that connects tasks, projects, learning, movement, books, habits and finance into one view of personal progress.
- shared event model across all modules
- works without a mandatory account and keeps data on-device
- architecture, Android, data, migrations and sequential releases
4ITX
Working HRTech platformProduct, ingestion and platform architecture
Current deep-tech vacancies, a multi-source data pipeline, shared taxonomy and explainable matching between professionals and roles.
- working vacancy and company catalog across five verticals
- data ingestion, cleaning, normalization and AI classification
- Talent Card, Talent Spec and explainable matching architecture
HOW WORK STARTS
A real task comes before a technical interview
The most useful evaluation starts with the problem, the current constraints and the required result. This is enough to expose engineering judgement, the quality of questions and the ability to turn ambiguity into a plan.
Task
You send the problem, current product or operational bottleneck.
System view
I identify unknowns, constraints, architecture boundaries and the first useful result.
Delivery format
We select architecture-only work, a focused pilot, a product milestone or full-cycle implementation.
Working result
I build, validate and deploy the agreed scope while retaining ownership of the engineering result.
ENGAGEMENT FORMATS
Three formats. One standard: the ability to own the result.
I consider project engagements, employment and partnerships selectively. The label itself is secondary: I assess the substance of the problem, the decision model, mutual obligations and whether my contribution can materially improve the product.
PROJECT ENGAGEMENT
Architecture, a complex milestone or full-cycle product delivery
I join a project when the business outcome, system boundaries and decision-maker can be made explicit. The engagement may cover architecture as a standalone deliverable, a focused pilot, a product milestone or complete implementation.
WHAT I ASSESS
- The problem benefits from a system-level approach
- Direct access to domain knowledge and the decision-maker
- The scope, timeframe and budget allow ownership of quality
EMPLOYMENT
A lead engineering or product role with real ownership
I consider employment when the role connects architecture, product decisions and delivery rather than isolating them into separate ticket queues. I am most effective where independent judgement is expected and the result matters more than process theatre.
WHAT I ASSESS
- Authority is proportionate to responsibility
- The role involves architecture, AI or full-cycle product delivery
- The team values autonomy and direct professional communication
PARTNERSHIP
Joint development of a product or technology direction
I enter a partnership when engineering is central to the product value and both sides are prepared to invest resources and make long-term decisions. The relationship requires transparent roles, economics and ownership from the outset.
WHAT I ASSESS
- Complementary expertise and a shared product ambition
- Agreed economics, decision rights and intellectual property
- Evidence that both sides are ready for sustained contribution
Have a problem that needs to become a working product?
Send the context, current constraints and desired outcome. I will outline the architecture, risks and the first useful milestone.