Middle AI-insener

Successfully completed an internship at Jusan Bank, acting as a Frontend Developer. Worked on projects including Umbrella, and developed ToDo List, covering the full development stack (front&back). Implementation of Python code using the Django framework, as well as client-side development on React JS. Also tested projects independently using case

Prompt engineering: designing roles, tones, contextual windows, toolbars; A/B testing of prompts and scenarios. Integrations: amoCRM, Bitrix24, Мой склад — pipelines, stages, fields, webhooks, automations, UTM bindings, manager interactions. Python for bots: webhook handlers, REST API, parsers, form validation, computational logic (prices/calculations/schedules), logging, and alerts. Channels: Telegram/WhatsApp/Instagram (connecting providers, buttons, menus, message templates). Analytics: events, tags, exports to tables/BI; lead/conversion reports. Client work: pre-project interviews, briefings, requirements formalization, assigning tasks to developers/designers, training, and documentation.

Developed AI-powered applications from scratch using Large Language Models (LLMs) and AI coding assistants within the Quadcode IDE. Architected and implemented RAG (Retrieval-Augmented Generation) pipelines — enabling knowledge-base search and contextually accurate AI assistant responses. Applied AI tools throughout the development lifecycle: backend logic generation, function writing, automated debugging, error analysis, and rapid MVP prototyping. Integrated third-party APIs and designed scalable system architectures using AI-assisted planning and code generation. Maintained and improved production code: developed new features, resolved bugs, and ensured stability across active projects. Gained hands-on expertise in LLM application development, prompt-to-code workflows, and AI-first software design principles.

Produced AI-generated video content and visual stories using Grok and Flow as core generation tools. Developed full creative pipelines from scratch: concept ideation →writing → storyboarding → AI scene generation → editing → final delivery. Designed and iterated prompts to precisely control visual style, character behaviour, atmosphere, and emotional tone across scenes. Built narrative structure for each video: scene sequencing, pacing, transitions, and visual storytelling logic. Delivered complete, publish-ready AI video productions and creative visual concepts for digital content channels.

Created and customized realistic AI avatars using Higgsfield. Developed avatar appearances, including styling, clothing, environments, and overall visual concepts. Created AI-powered videos with avatars using HeyGen. Worked withs, avatar movements, facial expressions, and presentation to achieve natural-looking content. Generated and tested multiple video variations to achieve the desired visual quality. Focused on realistic skin texture, lighting, facial details, and natural on-camera behavior.