{"id":3605389,"url":"https://alion.io/job/sibintek-senior-mlopsllmops-engineer","title":"Senior MLOps\\LLMOps Engineer","company":{"id":47320,"name":"Sibintek","domain":"sibintek.ru","url":"https://alion.io/company/sibintek","size_band":"1001-5000","is_staffing_agency":false,"employer_type":"direct","is_intermediary":false,"listed_via":null,"ats_vendor":"HeadHunter","truth_index":{"grade":"A","score":90,"open_postings":35,"ghost_share":0.114,"stale_share":0,"repost_share":0.229,"time_to_fill_p50_days":30,"computed_at":"2026-11-04T06:34:25Z"}},"role":"AI/ML","role_family":"AI/ML","seniority":"senior","employment_type":null,"work_mode":"remote","remote_scope":"stated_regions","remote_scope_basis":"board_field","remote_working_hours":null,"hiring_geo_confidence":"structured","locations":["Moscow, Russia"],"countries":["RU"],"hiring_countries":["RU","BY","KZ","KG","AM"],"hiring_countries_total":5,"salary":null,"salary_estimate":null,"experience_years_min":null,"visa_sponsorship":false,"relocation_package":false,"has_equity":false,"technologies":[{"name":"AWQ","optional":false},{"name":"C++","optional":false},{"name":"CUDA","optional":false},{"name":"DeepSpeed","optional":false},{"name":"GPTQ","optional":false},{"name":"Grafana","optional":false},{"name":"InfiniBand","optional":false},{"name":"Knative","optional":false},{"name":"KServe","optional":false},{"name":"Kubernetes","optional":false},{"name":"llama.cpp","optional":false},{"name":"LLM","optional":false},{"name":"Loki","optional":false},{"name":"LoRA","optional":false},{"name":"Mixture of Experts","optional":false},{"name":"NCCL","optional":false},{"name":"OpenTelemetry","optional":false},{"name":"Prometheus","optional":false},{"name":"Python","optional":false},{"name":"Ray","optional":false},{"name":"SGLang","optional":false},{"name":"TensorRT-LLM","optional":false},{"name":"TPOT","optional":false},{"name":"Triton","optional":false},{"name":"vLLM","optional":false},{"name":"FinOps","optional":true},{"name":"OCR","optional":true},{"name":"PEFT","optional":true},{"name":"Speech Recognition","optional":true},{"name":"TensorRT","optional":true},{"name":"Text-to-Speech","optional":true}],"status":"live","first_seen_at":"2026-11-02T11:34:07Z","employer_posted_date":"2026-11-02","last_verified_at":"2026-11-03T18:08:31Z","board_verified":true,"closed_at":null,"days_open":2,"trust":{"level":"ok","repost_count":null,"flags":[],"days_open":2},"description":"Обязанности: \n Разворачивать и эксплуатировать LLM-сервисы на GPU-кластере: инстансы vLLM/SGLang, тиринг по SLA, обновления без простоя парка;\n Тюнинговать движки: continuous batching, chunked prefill, prefix caching, планировщики, спекулятивное декодирование, квантование FP8/NVFP4;\n Строить мультинодовые схемы (TP/PP/EP) и P/D-дисагрегацию с передачей KV-кэша по RDMA;\n Настраивать наблюдаемость (TTFT/TPOT/P99, KV-hit, утилизация HBM);\n Участвовать в планировании мощности и токен-биллинге: бенчмарки, roofline-оценки, сценарии p10/p50/p99.\n Требования: Инференс-движки:\n Глубокое знание vLLM в проде: PagedAttention, prefix cache, планировщик и приоритеты, квантование, мультинод (Ray, TP/PP); умение читать исходники и патчить;\n Практика с одним из движков: SGLang, TensorRT-LLM, llama.cpp, DeepSpeed-MII/FastGen - и понимание компромиссов между ними;\n Квантование в проде: FP8, W8A8, GPTQ/AWQ; знание FP4/NVFP4/MXFP4 - механика, потери качества, применимость по тирам;\n Спекулятивное декодирование: как устроено, влияние на TPOT и пропускную способность, приём черновиков, когда вредит;\n Roofline-мышление: memory-bound декод против compute-bound префилла, роль пропускной способности HBM, footprint KV-кэша (GQA/MLA/сжатый KV) - умение прикинуть «влезет/не влезет» и ожидаемый throughput на узле;\n Кластер и сети:\n NVIDIA-стек: драйверы, CUDA, NCCL; диагностика (nvidia-smi, DCGM, чтение отчётов Nsight);\n RDMA/InfiniBand/RoCE: тесты ширины канала, латентность all-to-all, топология и тюнинг NCCL;\n MoE-инференс: экспертный параллелизм, all-to-all dispatch/combine (DeepEP-класс), горячие эксперты и балансировка;\n Платформа и инженерная база:\n Kubernetes в проде: GPU-шедулинг, device plugins/MIG, node pools, HPA/VPA;\n Деплой моделей: KServe/llm-d/Knative/Triton; реестры моделей, версионирование весов, LoRA-адаптация;\n Контейнеризация GPU-нагрузок: CUDA-образы, пиннинг потоков, hugepages;\n Наблюдаемость и надёжность:\n Prometheus/Grafana/Loki/OpenTelemetry: метрики SLA, трейсинг запроса через роутер, префикс-кэш и движок;\n Нагрузочное тестирование инференса: генераторы трафика по профилям вход/выход и cache-hit; умение отличить SLA-деградацию от ресурсной;\n Инцидент-практика: деградации P99, фрагментация памяти KV-кэша, OOM, отказы реплик.\n Языки:\n Python на продакшен-уровне (сервисы, автоматизация, бенчмарки); чтение C++/CUDA-кода движков;\n Go/Bash - для инструментария.\n Будет плюсом:\n Собственные бенчмарки, доклады или статьи по тюнингу инференса;\n Опыт P/D-дисагрегации и KV-пулов (NVIDIA Dynamo, Mooncake, DistServe);\n Мультиарендность: изоляция тенантов, защита TTFT от «толстых» промптов (VTC-подход);\n FinOps и токен-биллинг, планирование capacity;\n Контуры ASR/TTS/OCR, embedding/rerank-сервисы.\n Условия: работа в аккредитованной IT-компании (льготная ипотека и проч.);\n оформление по ТК РФ и его полное соблюдение (оплачиваемые отпуска, больничные);\n ДМС со стоматологией с первых недель работы.","description_format":"text","description_chars":2928,"description_truncated":false,"requirements":{"experience_years_min":null,"management_years_min":null,"team_size_min":null,"manages_managers":false,"education":null,"security_clearance":false,"languages":[{"language":"Russian","level":"Advanced (C1)","optional":false}]},"benefits":[],"hiring_locations":[{"name":"Russia","iso":"RU","kind":"country"},{"name":"Belarus","iso":"BY","kind":"region"},{"name":"Kazakhstan","iso":"KZ","kind":"region"},{"name":"Kyrgyzstan","iso":"KG","kind":"region"},{"name":"Armenia","iso":"AM","kind":"region"}],"hiring_excludes":[],"relocation_offered":false,"industries":["Managed IT Services (MSP)","Systems Integrators","Oil & Gas"],"lifecycle":[{"event":"open","at":"2026-11-02T11:37:59Z"}],"visa":[],"liveness":{"score":90,"band":"hot","label":"Hiring now","p_open":1,"p_active":0.903,"p_room":1,"age_days":1,"expected_fill_days":30,"reasons":["conf:12","velocity","win:early"],"computed_at":"2026-11-04T06:34:25Z"},"pay":null,"html_url":"https://alion.io/job/sibintek-senior-mlopsllmops-engineer","json_url":"https://alion.io/job/sibintek-senior-mlopsllmops-engineer.json","meta":{"generated_at":"2026-11-04T15:53:44Z","cache_seconds":300,"methodology":"https://alion.io/methodology","terms":"https://alion.io/terms","contact":"https://alion.io/contact","api":"https://alion.io/developers","about":"Alion is a live layer of people, companies and AI agents: who they are, whether they are real and active right now, what they do and how to work with them, readable by people and by agents and paid per call.","catalog":"https://alion.io/catalog.json","usage":{"tier":"crawler_verified","counted_by":"address","units_charged":1,"used_today":24306,"day_limit":null,"remaining_today":null,"minute_limit":300,"resets_at":"2026-11-05T00:00:00Z"}},"offers":[{"id":"company.slices","title":"One company in depth, by slice","status":"live","price":{"credits":0.02,"usd":0.002,"plus_per_slice":{"credits":0.05,"usd":0.005}},"unit":"per company, plus each slice with data","note":"the employer in depth","call":{"mcp_tool":"get_company","arguments":{"id":47320},"rest":"https://alion.io/mcp/rest/get_company?id=47320"},"human":"https://alion.io/catalog?offer=company.slices&for=job%2Fsibintek-senior-mlopsllmops-engineer"},{"id":"market.stats","title":"A market slice: pay, demand and time to fill","status":"live","price":{"credits":1,"usd":0.1},"unit":"per slice","note":"pay, demand and time to fill for this role and place","call":{"mcp_tool":"market_stats"},"human":"https://alion.io/catalog?offer=market.stats&for=job%2Fsibintek-senior-mlopsllmops-engineer"},{"id":"job.search","title":"Open jobs by role, technology, place, pay and visa","status":"live","price":{"credits":0.02,"usd":0.002},"unit":"per posting in a list","note":"similar open postings","call":{"mcp_tool":"search_jobs"},"human":"https://alion.io/catalog?offer=job.search&for=job%2Fsibintek-senior-mlopsllmops-engineer"},{"id":"company.verify","title":"Is this company real and active right now","status":"pilot","price":null,"unit":"per company","request":{"url":"https://alion.io/catalog/request","method":"POST","body":"{\"offer\": \"company.verify\", \"for\": \"job/sibintek-senior-mlopsllmops-engineer\", \"note\": \"what you need it for\"}"},"human":"https://alion.io/catalog?offer=company.verify&for=job%2Fsibintek-senior-mlopsllmops-engineer"}],"agent":{"mcp":"https://alion.io/mcp/jobs","mcp_open":"https://alion.io/mcp/open","next":[{"tool":"get_job","arguments":{"id":3605389}},{"tool":"search_jobs","arguments":{"company":"sibintek.ru","detail":"compact"}},{"tool":"match_people","arguments":{"job_id":3605389},"note":"people open to offers who fit this vacancy"}],"note":"The same data through MCP tools: https://alion.io/mcp/open needs no sign-in (compact searches, postings, people); https://alion.io/mcp with an Alion account has everything. Docs: https://alion.io/developers"}}