Senior Frontend Engineer
I build reliable, accessible, high-performance web applications.
Senior Frontend Engineer with 20 years of software development experience, specializing in enterprise applications built with React, Next.js, and TypeScript. I work across frontend architecture, design systems, performance optimization, accessibility, and technical leadership of teams of up to 10 engineers.
Iasi, Romania
- 20 years
- in software development
- Up to 10
- engineers led
- React · TypeScript · Next.js
- primary specialization
- Public safety · Healthcare · SaaS · AI
- delivery domains
In the terminal
Interfaces are not only for browsers
A text-mode dashboard: the same problems as any data-heavy UI — layout, refresh, legibility — with far tighter constraints.
What I do
Four areas I am relied on for
The work below recurs across every product I have delivered, from public safety dashboards to AI tooling.
Frontend architecture
Structuring long-lived applications so that feature teams can move without destabilizing the codebase.
- Component and module boundaries
- State strategy and data fetching
- Coding standards and reviews
Design systems and reusable components
Building typed, documented component libraries that keep product surfaces consistent across teams.
- TypeScript component libraries
- Storybook documentation
- Shared patterns across modules
Performance and real-time interfaces
Keeping data-heavy and streaming interfaces responsive through measurement and targeted optimization.
- Code splitting and lazy loading
- Bundle and rendering optimization
- RxJS backpressure for event streams
Accessibility and engineering leadership
Treating accessibility as a functional requirement, and growing the engineers who maintain it.
- Semantic HTML and keyboard operation
- WCAG-aligned review practices
- Mentoring, estimation, release planning
Applied AI
Working with AI, in production
AI work has been the centre of my role since 2025: not model training, but the engineering around it — the interfaces that let people configure a model-driven job, watch it run, and decide whether to trust what it produced.
LLM-driven workflows people can operate
An AI pipeline is only useful if someone can set it up, start it and follow it. That configure–execute–monitor loop is the work I owned at IBM.
- Source-code ingestion and project configuration
- Migration job creation and execution
- Monitoring a transformation while it runs
Asynchronous AI jobs as first-class UI state
Model-driven jobs run long and fail in more than one way. Progress, errors, retries and validation failures are modelled as expected states rather than error handling bolted on at the end.
- Explicit progress reporting during execution
- Retry and error paths surfaced in the interface
- Validation failures presented as reviewable results
Making generated output reviewable
The hardest screens in an AI product are the ones where a person decides whether the output is good enough. I built the review surfaces for exactly that decision.
- Transformation review interfaces
- Validation results
- Generated-code analysis
Cost and confidence made visible
Teams running LLM workloads need to see what a run is costing and how sure the system is, not just whether it finished.
- Token usage dashboards
- Confidence scores
- Migration progress reporting
Integrating AI orchestration services
The frontend sits on top of a Python orchestration layer and several data platforms, and has to surface what they report without leaking their internals into the interface.
- React and Next.js integrated with Python/FastAPI over REST
- Code parsing, LLM execution, validation and reporting
- Execution metadata from MCP, Snowflake and Databricks pipelines
Prompt engineering and persona design
Built independently: a platform where configurable personas, tone and conversation context drive what the model produces, with response quality as the product problem.
- Persona creation, configuration and management
- Replies generated from context, tone and persona characteristics
- Prompt engineering and response quality as the main focus
Formal grounding alongside the delivery work — 3 GlobalLogic AI certifications:
- AI 101: GL Certified AI Associate - Developers GlobalLogic
- Generative AI GlobalLogic
- Agentic AI with LangChain and Langflow GlobalLogic
Selected work
Engineering case studies
Four products, each with a different frontend problem at its centre.
- Enterprise AIIBM
Enterprise AI Code Transformation Platform
Frontend for a platform that migrates applications between programming languages using AI-assisted transformation, with configuration, execution and review workflows.
- AI-assisted migration workflows
- Asynchronous job monitoring
- Generated-code review
- Public safetyAROBS Transilvania Software
Nationwide Disaster Alert Platform
Mission-critical React and Vue.js interfaces for a nationwide emergency alert platform rendering high-volume real-time event streams.
- Real-time event streams
- RxJS backpressure
- TypeScript component library
- HealthcareAROBS Transilvania Software
Telehealth Product
Next.js application for appointment scheduling and video consultations, with a shared Storybook component library for patient and provider modules.
- Next.js and server-side rendering
- Video consultation workflows
- Storybook component library
- SaaSAROBS Transilvania Software
Multi-tenant Booking SaaS
React and Next.js frontend for a multi-tenant booking and appointment platform, with Apollo Client GraphQL caching and a shared Storybook library.
- Multi-tenant frontend
- GraphQL with Apollo Client
- Shared Storybook components
Built independently
Personal projects
Work outside client and employer projects — publicly available, so you can try it rather than take my word for it.
- Personal projectpersonareply.com
PersonaReply (opens in a new tab)
AI-generated replies driven by configurable personas
An AI-powered platform that helps users generate personalized, context-aware replies using configurable personas. Users define different communication styles, personalities and tones, which lets the AI produce responses adapted to a specific conversation and audience.
- Designed an intuitive dashboard for creating, configuring and managing AI personas.
- Built workflows for generating relevant replies based on conversation context, selected tone and persona characteristics.
- Integrated generative AI capabilities to deliver natural, personalized and consistent responses.
- Focused on prompt engineering, response quality, usability and a clean, responsive user experience.
- Structured the application to support additional personas, communication scenarios and AI models as the platform evolves.
A practical use of generative AI to make digital communication faster, more personalized and more effective.
- Generative AI
- Persona configuration
- Prompt engineering
- Responsive UI
Career
From freelance in 2006 to enterprise AI tooling
Twenty years of software development, the last decade focused on enterprise React and TypeScript.
Senior Full Stack Developer
IBM
Bucharest, Romania (remote)
Frontend-focused role on an enterprise AI-driven code transformation platform that migrated applications between programming languages while preserving business logic.
Senior Frontend Developer
AROBS Transilvania Software
Iasi, Romania
Ten years of frontend delivery across four long-running products in public safety, healthcare, hospitality SaaS and contract management.
Frontend Developer
Conex Group
Iasi, Romania
Internal portal for more than 500 employees.
Frontend Developer
Centric IT Solutions Romania
Iasi, Romania
Public-sector and social-care digital services for a Swiss client.
JavaScript Developer (freelance)
Crismaru Constantin PFA
Romania
Responsive websites and custom web solutions for small and medium-sized businesses.
Toolkit
Skills, grouped by what they are for
Grouped by capability rather than listed as a keyword cloud.
Frontend engineering
The core stack used daily across enterprise product work.
- React
- Next.js
- TypeScript
- JavaScript ES6+
- React Hooks
- React Router
- Redux
- RxJS
- HTML5
- CSS3
- Sass
Architecture and performance
How applications are structured and kept fast as they grow.
- Frontend architecture
- Component architecture
- Design systems
- Storybook
- Server-side rendering
- Code splitting
- Lazy loading
- Rendering and bundle optimization
- Accessibility and WCAG practices
Testing and quality
Test tooling used to keep behavior verifiable during change.
- Playwright
- React Testing Library
- Jest
- Cypress
APIs and real-time integration
Connecting interfaces to services, streams and AI workflows.
- REST APIs
- GraphQL
- Apollo Client
- WebRTC
- Real-time event streams
- Kafka
- RabbitMQ
Backend and data
Supporting full-stack capability alongside the frontend specialization.
- Node.js
- NestJS
- Express.js
- Python
- FastAPI
- PostgreSQL
Cloud and delivery
Build, deployment and collaboration tooling.
- Docker
- AWS
- Azure
- Webpack
- Git
- GitLab
- Agile
- Scrum
Leadership
Technical leadership
Alongside hands-on delivery, I have led frontend teams and owned the standards they work to.
Led teams of 8-10 engineers
Coordinated a frontend team on a contracts management platform, from estimation through release delivery.
Frontend architecture ownership
Owned architecture decisions and component design across React, Next.js, Vue.js and AngularJS codebases.
Estimation and release planning
Planned scope and coordinated releases with product and backend teams in Agile and Scrum settings.
Code review and mentoring
Mentored engineers through code review and architecture discussions, and established shared frontend development standards.
Have a senior frontend, React, or full-stack opportunity?
Send a message and include the role, location model, and expected start date.