Jobs at Prosus Group

Build and operate secure, scalable backend services for Toqan, Prosus’s internal AI platform. Responsibilities include developing agent runtimes, orchestrating LLMs and tools, implementing asynchronous workflows, improving reliability and observability, integrating services, making architectural decisions, and collaborating with product, platform, and AI teams. The role primarily uses Go and AWS and involves CI/CD, automated testing, production operations, and hands-on development of agentic AI capabilities.

Own the multi-year strategy and roadmap for Prosus's AI platform, including agent infrastructure, evaluation, deployment, and shared capabilities. Lead and mentor product managers, partner with engineering and applied science on technical trade-offs, align global stakeholders, and drive platform adoption, reliability, and measurable business impact. The role requires deep AI and platform expertise, strong product leadership, comfort with ambiguity, and experience operating at scale.

Develop and deploy domain-specific language models through pre-training, fine-tuning, and post-training. Build distributed multi-node GPU training infrastructure, prepare large-scale datasets, generate synthetic data, optimize training stability and throughput, and create evaluation frameworks. Apply reinforcement learning techniques such as RLHF, DPO, and GRPO while translating research into reliable production systems used by millions of users.

Build examples, integrations, tooling, and documentation for Toqan, Prosus’s agentic AI platform. Reproduce bugs, contribute fixes, support developers, review pull requests, and convert community feedback into engineering improvements. Create technical content, demos, talks, workshops, and events to grow adoption. The role requires production software engineering experience, hands-on agent and LLM expertise, MCP familiarity, cloud experience, and strong public technical communication skills.

Build and configure AI-powered media monitoring, stakeholder CRM, KPI dashboards, and automated reporting for Group Communications. Clean and enrich contact data, define monitoring alerts, produce media insights, train users, support GDPR compliance, and document systems for handover. Work directly with senior stakeholders to improve communications decision-making through AI tools, data, and automation.

Leads Prosus’s product communications function for AI products, owning positioning, messaging, launches, media and analyst relationships, owned channels, and cross-functional alignment. Builds the communications discipline through playbooks, processes, measurement, and standards. Translates complex technology into compelling narratives for technical and non-technical audiences while connecting communications to adoption, reputation, and business results.

Owns Prosus’s global brand experience strategy across flagship and external technology events. Leads event concepts, physical brand expression, audience experience, activations, production, budgets, vendors, agencies, and execution across international markets. Partners with programming, content, speaker, brand, and communications teams to create distinctive experiences for executives, investors, policymakers, and industry audiences while developing repeatable standards and toolkits.

Conduct research and build experimental systems using LLMs, causal inference, experimentation, and optimization. Develop treatment-effect, uplift, and counterfactual models; analyze randomized trials; and create constrained optimization models for promotions and incentives. Read research papers, prototype ideas, and translate successful approaches into production systems. The internship requires collaboration in Amsterdam for 6–12 months, with at least three office days weekly.

Leads technical security assessments, cyber risk audits, vulnerability investigations, and architectural reviews across global businesses. Reports findings to CTOs and CISOs, translates technical risks into business recommendations, coordinates high-stakes security initiatives, and supports technical experts. The role also builds the global cybersecurity community, organizes events, participates in hackathons, and investigates emerging AI, ML, LLM, and agent security risks. Travel of up to 30% is required.

Leads the training and deployment of domain-specific language models at scale. Responsibilities include designing experiments, improving model quality and efficiency, building distributed multi-node GPU infrastructure, preparing 100B+ token datasets, generating synthetic data, optimizing inference, creating evaluation frameworks, debugging training stability, and mentoring ML engineers and interns. The role requires production-grade Python and PyTorch expertise, technical leadership, and hands-on experience with large-scale language-model training and deployment.

Train, fine-tune, evaluate, and serve open-source large language models such as Qwen and Mistral. Design experiments, implement research findings, and contribute across the machine learning lifecycle, including production serving infrastructure. The role involves Python, PyTorch, Hugging Face, model training, data processing, distributed systems, and collaboration with AI engineers. This is a 6–12 month internship based in Amsterdam with at least three office days weekly.