IT Craft — вакансії

IT Craft invites an AI/ML Python Developer to join the company.

Requirements:

• Core Software Engineering: 5 + years Python 3.x; strong proficiency with FastAPI, Django or Flask; Git, unit testing, CI/CD, Docker;
• Generative AI & LLMs: Practical knowledge of leading models (GPT, Claude, Gemini, Mistral, Llama, Qwen, etc); working with them via public APIs, self-hosting or managed platforms (AWS Bedrock, Azure OpenAI, GCP Vertex);
• LLM Techniques: Retrieval-Augmented Generation (RAG), Supervised Fine-Tuning (SFT), Parameter-Efficient Fine-Tuning (LoRA/QLoRA/PEFT), Reinforcement Learning from Human Feedback (RLHF), prompt engineering and evaluation;
• Chatbots & Agents: Designing multi-turn conversational flows, tool invocation, agent routing/orchestration with LangChain, LlamaIndex, CrewAI, Semantic Kernel or similar;
• Vector Databases: Experience with FAISS, Milvus, Pinecone, Weaviate, Chroma, pgvector (indexing, similarity search, hybrid search);
• MCP Servers: Building and integrating Model Context Protocol (MCP) servers to securely expose domain tools/functions to LLMs;
• Automation Platforms: Advanced use of no-/low-code tools: n8n, Make (Integromat), Zapier (triggers, webhooks, custom modules);
• Soft Skills: Ownership mindset, rapid prototyping, clear technical writing; English at Upper-Intermediate (B2) or higher.

Key Responsibilities:

• Design & build end-to-end LLM workflows in Python (FastAPI, Django or Flask) — from data ingestion to inference APIs;
• Develop and fine-tune models (SFT, LoRA/QLoRA, RLHF) using PyTorch and modern open-source tooling;
• Implement RAG pipelines: document parsing, embedding generation, vector search and evaluation;
• Create chatbots / autonomous agents with frameworks such as LangChain, LlamaIndex, Semantic Kernel or CrewAI;
• Expose tools via MCP servers so that external LLMs can invoke domain-specific actions;
• Integrate no-/low-code automation (n8n, Make, Zapier) to orchestrate business workflows;
• Write clean, test-driven code; set up CI/CD; monitor, log, and continuously improve model performance and system reliability.

Nice to Have:

• Experience with data privacy & compliance (GDPR, SOC 2);
• Operate in cloud & on-prem (Docker, GPU instances, AWS Bedrock/Azure OpenAI, self-hosted LLMs, etc.);
• Hands-on experience with TensorFlow, PyTorch, NumPy, Pandas; ability to train, fine-tune, and evaluate ML/LLM models;
• Familiarity with front-end stacks (React/Next.js, TypeScript) for building AI-powered user interfaces;
• Bonus: Experience in Generative Graphics, including Stable Diffusion, Inpainting techniques, and ControlNets;
• Bonus: Experience in LLM hosting and working with llama.cpp.