Want to know if this job is worth applying to?
София, София-град, България
Hybrid
Welcome to TELUS Digital — where innovation drives impact at a global scale. As an award-winning digital product consultancy and the digital division of TELUS, one of Canada’s largest telecommunications providers, we design and deliver transformative customer experiences through cutting-edge technology, agile thinking, and a people-first culture.
With a global team across North America, South America, Central America, Europe, Africa, and APAC, we offer end-to-end expertise across various service offerings: Web, Mobile & Digital Marketing | Enterprise AI | Customer Care AI & Technology | Enterprise Technology Modernization
From mobile apps and websites to voice UI, chatbots, AI, customer service, and in-store solutions, TELUS Digital enables seamless, trusted, and digitally powered experiences that meet customers wherever they are — all backed by the secure infrastructure and scale of our multi-billion-dollar parent company.
Our Senior Full-Stack AI Engineers are integral parts of our team at TELUS Digital. To help retain our deep culture of collaboration, this role will maintain an in-office presence in a hybrid capacity in Sofia, Bulgaria
Business-level English proficiency (written and verbal) is required
As a Senior Full-Stack AI Engineer to help build and evolve the AI systems that power a leading gaming client's player support platform. This role sits at the center of two connected products: an AI assistant used directly by players, and an AI copilot used by the support agents who help them. You'll work across the full stack — from LLM orchestration and agentic workflow design on the backend to the React applications and admin tooling that let the team configure, monitor, and improve these systems in production.
This is a hands-on engineering role for someone who has already shipped and operated production AI systems, not just prototyped them. You'll design conversation flows that handle ambiguous intent, tool use, retrieval, and escalation to a human; evaluate and tune models across multiple providers; and build the observability and resilience that keep these systems reliable at scale. Day-to-day, you'll work primarily in TypeScript across a Node.js backend and React frontend, using state machines to model complex conversational workflows and integrating with large language model providers such as OpenAI, Anthropic, and Gemini.
AI Platform & Agent Systems
Design and extend production-grade LLM applications and agentic workflows, including RAG, intent detection, clarification, fulfillment, escalation, and human-in-the-loop flows
Build and maintain the underlying conversation engine, including flow validation, configuration tooling, and the admin experience used to manage it
Integrate tool-use and agentic protocols to extend what the AI assistant and copilot can do
Evaluate, benchmark, and tune models across multiple providers, owning model selection decisions that balance quality, latency, throughput, reliability, and cost
Troubleshoot production AI issues such as hallucinations, retrieval failures, prompt regressions, model drift, and provider outages
Build resilience mechanisms including retries, fallback routing, caching, streaming, and rate limiting
Instrument and tune model quality using tracing, evaluation datasets, A/B testing, and prompt versioning
Frontend & Admin Tooling
Build player-facing and support-agent-facing experiences in React
Develop and maintain the admin interface used to manage flows, intents, releases, translations, prompts, and evaluations
Integrate real-time, event-driven product features
Own frontend performance, accessibility, and internationalization
Platform & Quality
Write comprehensive automated tests across unit, integration, contract, and end-to-end layers
Build observability for both traditional services and AI systems, including tracing, dashboards, and alerting
Diagnose performance bottlenecks across application code, retrieval systems, model inference, and infrastructure
Participate in an on-call rotation and respond to production incidents with clear communication
Document architectural decisions and keep runbooks current
Proven experience building and operating production LLM-powered systems — chatbots, AI assistants, agent copilots, RAG systems, or LLM orchestration platforms
Strong TypeScript/Node.js engineering skills, with fluency in strict-mode TypeScript
Hands-on production AI experience: prompt engineering, RAG pipelines, agent design, tool calling, model evaluation, observability, and failure-mode analysis
Full-stack depth — comfortable moving between backend APIs, React UIs, databases, infrastructure, and production operations
Ability to evaluate tradeoffs between model quality, latency, reliability, throughput, and cost
Ability to troubleshoot AI systems across prompts, retrieval pipelines, model configuration, infrastructure, and application code
Experience modeling complex asynchronous workflows (state machine experience is a strong signal)
Solid understanding of REST API design, asynchronous patterns, and caching strategies
A strong testing culture, with unit, integration, and contract tests treated as first-class deliverables
Experience working in a monorepo with multiple interconnected services
Experience with agentic tool-use protocols (e.g., MCP), semantic search/retrieval architectures, or AI gateways
Familiarity with LLM observability and evaluation platforms, and experience evaluating multiple foundation models or providers at scale
Experience building AI copilots, assistants, or other conversational products, especially under low-latency or streaming constraints
Hands-on experience with NestJS specifically
Background in customer support or player support platforms, gaming, or high-volume consumer products
At TELUS Digital, we are proud to be an equal opportunity employer and are committed to creating a diverse and inclusive workplace. All aspects of employment, including the decision to hire and promote, are based on applicants’ qualifications, merits, competence and performance without regard to any characteristic related to diversity.
We will only use the information you provide to process your application and to produce tracking statistics. Since we do not request personal data deemed sensitive, we ask you to abstain from sharing that information with us.
For more information on how we use your information, see our Privacy Policy.
