AI Engineer - Gaming Analytics Prototypes

About the role

Sensor Tower's digital intelligence platform gives insight into market trends, advertising, competitors' strategies and user audiences across mobile, web and gaming. Its mobile gaming portfolio includes AppMagic and Playliner, which help game developers and publishers understand the market and how top games run LiveOps. The company is looking for an AI Engineer to build LLM-powered prototypes for its gaming analytics products, turning product ideas into working prototypes on real data and taking the successful ones to production with the engineering teams.

Key responsibilities

• Build AI prototypes for game analytics, focusing on multimodal data such as screenshots, images and gameplay video, for example recognizing LiveOps events from in-game screenshots or detecting changes in a game update.
• Choose and use the right models for each task (mainly Claude), balancing quality, speed and cost.
• Build data pipelines with the Data Engineer to run models on large volumes of game data.
• Measure quality so the team knows what works and what does not.
• Build retrieval and multi-step pipelines.
• Turn the best prototypes into real product features together with the engineers.

Required profile

• Strong Python; TypeScript is a plus.
• Hands-on experience building products on top of LLM APIs, not just chatbots.
• Understanding of RAG and vector search, and when they are not needed.
• Able to measure results and show how well a solution works.
• Can work independently and explain decisions clearly.
• Ability to pass a technical interview in both Russian and English.

Required skills

• Experience with multimodal models (screenshots, images or video).
• Preferred: interest in or experience with mobile games, LiveOps or market analytics.
• Preferred: hands-on experience in building AI projects.
• Preferred: experience with LangGraph, LlamaIndex or similar frameworks and an opinion on when they are worth using.

Candidates apply through the company website or through posts linked to the company profile on sites such as LinkedIn, Indeed, BuiltIn or Hackajob. Only candidates who apply through these channels will be considered.