{"id":2630304,"url":"https://alion.io/job/tatarstan-information-technology-center-ml-engineer-llm","title":"ML Engineer (LLM)","company":{"id":3817081,"name":"Tatarstan Information Technology Center","domain":"citrt.ru","url":"https://alion.io/company/tsentr-informatsionnykh-tekhnologii-rt","size_band":"51-200","is_staffing_agency":false,"employer_type":"direct","is_intermediary":false,"listed_via":null,"ats_vendor":"HeadHunter","truth_index":null},"role":"AI/ML","role_family":"AI/ML","seniority":null,"employment_type":null,"work_mode":"hybrid","remote_scope":null,"remote_scope_basis":null,"remote_working_hours":null,"hiring_geo_confidence":"structured","locations":["Kazan, Russia"],"countries":["RU"],"hiring_countries":[],"hiring_countries_total":0,"salary":null,"salary_estimate":null,"experience_years_min":null,"visa_sponsorship":false,"relocation_package":false,"has_equity":false,"technologies":[{"name":"AI Agents","optional":false},{"name":"AWQ","optional":false},{"name":"DeepSeek","optional":false},{"name":"DPO","optional":false},{"name":"Embeddings","optional":false},{"name":"FAISS","optional":false},{"name":"Fine-tuning","optional":false},{"name":"Flash Attention","optional":false},{"name":"Function Calling","optional":false},{"name":"GGUF","optional":false},{"name":"GPTQ","optional":false},{"name":"Hugging Face","optional":false},{"name":"LLM","optional":false},{"name":"LoRA","optional":false},{"name":"Machine Learning","optional":false},{"name":"Milvus","optional":false},{"name":"Mistral","optional":false},{"name":"Mixture of Experts","optional":false},{"name":"Multimodal AI","optional":false},{"name":"NER","optional":false},{"name":"NLP","optional":false},{"name":"NVLink","optional":false},{"name":"OpenAI","optional":false},{"name":"Perplexity","optional":false},{"name":"Qdrant","optional":false},{"name":"QLoRA","optional":false},{"name":"Qwen","optional":false},{"name":"RAG","optional":false},{"name":"RLHF","optional":false},{"name":"Semantic Search","optional":false},{"name":"Semantic Search","optional":false},{"name":"Sentence-Transformers","optional":false},{"name":"SFT","optional":false},{"name":"SGLang","optional":false},{"name":"Speculative Decoding","optional":false},{"name":"TGI","optional":false},{"name":"Transformers","optional":false},{"name":"vLLM","optional":false},{"name":"Weaviate","optional":false},{"name":"AudioCraft","optional":true},{"name":"CLIP","optional":true},{"name":"CogVideoX","optional":true},{"name":"ControlNet","optional":true},{"name":"DreamBooth","optional":true},{"name":"HunyuanVideo","optional":true},{"name":"LCM","optional":true},{"name":"Model Distillation","optional":true},{"name":"MusicGen","optional":true},{"name":"PEFT","optional":true},{"name":"SDXL","optional":true},{"name":"Stable Diffusion","optional":true},{"name":"xFormers","optional":true}],"status":"live","first_seen_at":"2026-10-15T15:34:08Z","employer_posted_date":null,"last_verified_at":"2026-10-15T15:34:08Z","board_verified":false,"closed_at":null,"days_open":3,"trust":{"level":"not_scored","repost_count":null,"flags":[],"days_open":3},"description":"Обязанности: \n Деплой, конфигурирование и поддержка LLM-сервисов в продакшне (vLLM, SGLang, TGI и др.).\n Оптимизация inference-пайплайнов для высоконагруженных систем (10 000+ одновременных пользователей).\n Квантизация моделей (GPTQ, AWQ, GGUF, SmoothQuant и др.).\n Fine-tuning и дообучение LLM (SFT, LoRA/QLoRA, RLHF/DPO).\n Подбор моделей, проведение экспериментов и исследований в сфере LLM.\n Разработка решений на основе LLM: RAG, агенты, function calling и др.\n Улучшение существующих решений и мониторинг качества моделей в промышленной эксплуатации.\nТребования: \n Уверенное знание LLM-serving фреймворков: vLLM, SGLang и других фреймворков для Text Generation Inference- понимание их архитектурных отличий, сценариев применения, преимуществ и ограничений\n Глубокое понимание механизмов оптимизации инференса: continuous batching, PagedAttention, speculative decoding, flash attention, tensor parallelism, pipeline parallelism\n Опыт квантизации LLM: GPTQ, AWQ, GGUF, INT4/INT8, SmoothQuant\n Опыт дообучения LLM: SFT, LoRA / QLoRA, DPO, RLHF - понимание когда и что применять\n Работа с популярными open-source моделями (LLaMA, Mistral, Qwen, DeepSeek, GPT OSS) и их весами (HuggingFace Transformers)\n Понимание и практический опыт построения RAG-систем, цепочек промптов, agentic-пайплайнов\n Опыт работы с высоконагруженными системами: оценка throughput/latency, профилирование, горизонтальное масштабирование, балансировка запросов\n Базовое понимание аппаратной части: GPU-архитектуры (NVIDIA A100/H100/B200), VRAM management, NVLink, multi-GPU setups\n Понимание и изучение SOTA-решений в сфере LLM для последующей имплементации в проект\n Знания базовых алгоритмов машинного обучения и основ трансформерных архитектур (attention, MoE, RoPE и пр.)\n Знание OpenAI-совместимых API\n Опыт с multimodal LLM (Kimi K2.5, Qwen-VL и пр.)\n Знание классических и современных NLP-задач: NER, text classification, summarization, machine translation, question answering, text generation\n Понимание этапов NLP-пайплайна: токенизация, нормализация, работа с BPE/SentencePiece/tiktoken\n Опыт работы с embeddings-моделями (sentence-transformers, E5, BGE и пр.) и векторными хранилищами (Weaviate, Qdrant, Milvus, FAISS) в контексте RAG и semantic search\n Понимание метрик качества NLP-моделей: BLEU, ROUGE, BERTScore, perplexity, human eval\n Будет плюсом:\n Понимание архитектур диффузионных моделей: DDPM, DDIM, LDM (Latent Diffusion), flow matching\n Опыт работы с text-to-image моделями (Stable Diffusion, FLUX, Kandinsky и пр.) и их компонентами - VAE, UNet/DiT, CLIP/text encoder\n Базовое понимание video generation моделей (Wan, CogVideoX, HunyuanVideo, Mochi и пр.) и их отличий от image-моделей\n Базовое понимание audio generation моделей (AudioCraft/MusicGen, Stable Audio, Bark и пр.)\n Опыт fine-tuning диффузионных моделей: DreamBooth, LoRA, ControlNet, IP-Adapter\n Знание inference-оптимизаций для диффузионок: SDXL Turbo / LCM / Lightning (distillation-подходы), cfg-distillation, tiling, xformers, torch.compile\n Опыт с диффузионными, image/video/audio generation моделями и их fine-tuning/оптимизацией.\n Условия:Официальное трудоустройство в соответствии с ТК РФ;\n Стабильная «белая» заработная плата, выплаты два раза в месяц (5 и 20 число), уровень зарплаты обсуждается по итогам интервью;\n Комфортный офис в центре города (новый айти-парк);\n График работы, 5/2;\n Поддержка профессионального развития;\n Участие в конференциях и обучающих мероприятиях за счет компании;\n Демократичный, молодой коллектив.\n Неформальные встречи и корпоративные мероприятия.\n Бронирование через военкомат при наличии военного билета (для военнообязанных) и айти-отсрочка.\n Дресс-код casual","description_format":"text","description_chars":3670,"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":[]},"benefits":[],"hiring_locations":[{"name":"Russia","iso":"RU","kind":"country"}],"hiring_excludes":[],"relocation_offered":false,"industries":["Cybersecurity"],"lifecycle":[{"event":"open","at":"2026-10-15T15:59:31Z"}],"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":33,"reasons":["seen:1","velocity","win:early"],"computed_at":"2026-10-17T05:46:00Z"},"pay":null,"html_url":"https://alion.io/job/tatarstan-information-technology-center-ml-engineer-llm","json_url":"https://alion.io/job/tatarstan-information-technology-center-ml-engineer-llm.json","meta":{"generated_at":"2026-10-18T15:43:12Z","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":21640,"day_limit":null,"remaining_today":null,"minute_limit":300,"resets_at":"2026-10-19T00:00:00Z"}},"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":3817081},"rest":"https://alion.io/mcp/rest/get_company?id=3817081"},"human":"https://alion.io/catalog?offer=company.slices&for=job%2Ftatarstan-information-technology-center-ml-engineer-llm"},{"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%2Ftatarstan-information-technology-center-ml-engineer-llm"},{"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%2Ftatarstan-information-technology-center-ml-engineer-llm"},{"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/tatarstan-information-technology-center-ml-engineer-llm\", \"note\": \"what you need it for\"}"},"human":"https://alion.io/catalog?offer=company.verify&for=job%2Ftatarstan-information-technology-center-ml-engineer-llm"}]}