Senior ML Engineer

📍 Almaty

🌍 We are looking for a Senior ML Engineer who can independently handle complex tasks from problem definition to production: design solutions, write code, set up pipelines, and monitor the quality of results. You will work on a wide range of AI areas — from classic ML and CV to RAG systems, LLM agents, and audio/speech technologies. The ML engineer uses ready-made and pre-trained models as building blocks; new model development is directed by Data Science.

🧵 Responsibilities:

• Independent development of AI applications: from prototype to production service
• Design and implementation of RAG systems, LLM agents, CV, audio, and ML pipelines
• Integration and deployment of pre-trained models and LLM providers; defining requirements for Data Science, accepting artifacts
• Development and support of REST API on FastAPI / Flask for AI services
• Writing and supporting Airflow DAGs; working with data at all stages of the pipeline
• Monitoring model quality metrics in production, participating in degradation analysis
• Code review, participation in architectural discussions, assisting colleagues' growth

🔊 Requirements:

• Experience in developing ML/AI systems for over 2 years
• Proficient in Python; practical experience developing services with FastAPI
• Experience with classical ML algorithms (scoring, fraud, recommendations): sklearn, XGBoost, CatBoost
• Practical experience with CV: integration and deployment of pre-trained detection, classification, or segmentation models (YOLO, SAM, OpenCV)
• Experience building RAG pipelines on LangChain and working with vector databases (Milvus)
• Understanding of LLM agent principles: tool use, prompt engineering, orchestration
• Practical experience with audio/speech: ASR (Whisper, NeMo, Triton), TTS (Piper, CosyVoice)
• Depth in at least two of the four areas (classic ML / CV / RAG and LLM agents / audio-speech) with a willingness to learn the others
• Experience with Airflow: writing DAGs, debugging, dependency management
• Understanding of MLOps practices: versioning, monitoring, experiment reproducibility

🔉 Will be a plus:

• Depth in additional areas out of the four
• Embeddings: sentence-transformers, fastembed
• Kafka, Celery + Redis
• Serving: NVIDIA Triton, vLLM; MLflow, DVC
• Platform: OpenShift, ArgoCD and GitOps, OpenDataHub
• Observability: Prometheus, Grafana, OpenTelemetry
• Dify, Streamlit

📩 You can contact me via private message or email: aruzhan.esimzhanova@bcc.kz

#ITvacancy #fintech #Almaty #fintech #ML #MLOps #AI #NLP