Работал в 3 компаниях 9 лет 2 месяца
IT
AI Engineer
Название компании скрыто
IT
4 года 4 месяца
08.2021 - 11.2025
- Owned GenAI delivery. From idea to production.
- Built LLM-powered features: chat, summarization, extraction, assistants.
- Implemented RAG over internal data. Grounded answers.
- Built ingestion pipelines: parse > chunk > embed > index.
- Added vector search + hybrid retrieval (keyword + vector).
- Designed tool/function calling. Deterministic workflows.
- Enforced structured outputs (JSON schema). Reduced hallucinations.
- Built evaluation suite: golden tests, prompt regression, failure replay.
- Implemented guardrails: prompt-injection defenses, PII handling, policy checks.
- Optimized cost/latency: caching, batching, model routing, rate limits.
- Deployed services with Docker/K8s. Production SLAs.
- Added observability: traces, metrics, logs for AI flows.
- Led architecture reviews. Mentored engineers. Owned standards.
AI Engineer
Название компании скрыто
IT
3 года 4 месяца
04.2018 - 07.2021
- Delivered AI features for multiple client products. Production mindset.
- Built backend APIs for AI workflows. Clean contracts.
- Implemented document intelligence: extraction, classification, normalization.
- Built search assistants: knowledge base Q&A with retrieval + citations.
- Integrated LLM APIs and tool workflows into existing systems.
- Built async processing: queues, workers, retries, idempotency.
- Productionized AI endpoints: auth, rate limits, timeouts, fallbacks.
- Implemented monitoring + alerting for model/API failures.
- Tuned performance: lower response times, stable throughput.
- Supported releases and hotfixes. Calm under pressure.
- Worked cross-team with product/QA/frontend. No integration surprises.
AI Engineer
Название компании скрыто
IT
1 год 6 месяцев
10.2016 - 03.2018
- Built backend foundations for data-driven features.
- Implemented data pipelines: cleaning, normalization, scheduled processing.
- Built rule-based automation > later upgraded to ML/AI-ready flows.
- Designed reliable storage and APIs for analytics/search.
- Improved stability and release discipline.
- Collaborated daily with QA/design/backend teams. Delivered on time.
Ключевая информация
- GenAI / LLM engineering
- LLM APIs (OpenAI/Claude/Gemini), prompt engineering, tool/function calling
- Structured outputs (JSON schema), deterministic workflows, fallback strategies
- RAG: chunking, embeddings, retrieval, reranking, grounding, citations
- Agents/workflows: multi-step tool use, orchestration, state handling
- Evaluation: golden datasets, prompt regression, quality gates, human review loops
- Safety: prompt-injection defenses, PII handling, access controls, content filtering
- Data & vector search
- ETL/ELT pipelines, data cleaning, document parsing (PDF/HTML/CSV)
- Vector DB/search (pgvector/Pinecone/Weaviate/FAISS), hybrid search patterns
- Indexing strategies, incremental updates, deduplication, versioned corpora
- AI Backend
- API design (REST/gRPC), microservices, async jobs/queues, caching (Redis)
- Reliability: retries, idempotency, circuit breakers, rate limiting
- Security: OAuth2/JWT, RBAC, secrets management, audit logging
- MLOps / LLMOps
- Model serving patterns (vLLM/TGI or managed inference), autoscaling
- CI/CD for AI services, environment parity, safe rollouts, rollback plans
- Observability: OpenTelemetry, structured logs, AI latency/cost dashboards
- Cloud & delivery
- Docker, Kubernetes, GitHub Actions/GitLab CI
- AWS/GCP/Azure basics (compute, storage, secrets, monitoring)
Учился в 1 заведении
Alte University
Bachelor of Computer Science
2018
Владеет языками
Английский
родной
Может проходить собеседование на этом языке
Может проходить собеседование на этом языке
Анонимный соискатель
AI Engineer
Киев
Голосеевский район, Дарницкий район, Деснянский район, Днепровский район, Оболонский район, Печерский район, Подольский район, Святошинский район, Соломенский район, Шевченковский район
Активно ищет работу
полная занятость, неполная занятость, проектная работа
Характер работы: удаленная работа
Обновлено 2 недели назад