AI Platform Engineer

About Tandems

Tandems builds intelligent solutions that reimagine how advice is delivered, how institutions operate, and how expertise is unlocked at scale. Our AI-enabled platform empowers advisors, bankers, and agents with deeper insights, personalized recommendations, and more time to focus on guiding clients through life's most important decisions. Backed by Bain Capital Ventures, General Atlantic, Wells Fargo, UBS, Santander, New York Life, and Comerica, with $100M+ in funding.

As an AI Platform Engineer, you'll design and develop the foundational capabilities that make our AI-powered products possible — from retrieval and orchestration layers to model integration and evaluation frameworks. You'll work across the stack to turn LLM capabilities into production-ready, compliant, and reliable features.

How You'll Make an Impact:

• Design and build the core AI platform services that power advisor-facing features such as intelligent insights, meeting summaries, and next-step recommendations.
• Develop and maintain retrieval-augmented generation (RAG) and semantic search pipelines.
• Integrate foundation models (OpenAI, Anthropic, AWS Bedrock, Azure OpenAI, or open-source alternatives) into scalable microservices and APIs.
• Create and manage prompt orchestration and versioning frameworks.
• Develop evaluation, monitoring, and observability systems to ensure AI quality.
• Optimize inference performance, latency, and cost through model routing, caching, and fine-tuning strategies.

Ideal Person for this Role: Core Engineering Skills:

• 4+ years of software engineering experience with production systems
• Strong proficiency in Python and hands-on experience with LLM frameworks (LangChain, LlamaIndex, Semantic Kernel, or DSPy)
• Proven experience building agent workflows and orchestrating multi-step AI processes
• Deep understanding of foundation model APIs

Agent Development Experience:

• Building autonomous and semi-autonomous agent systems that can plan, use tools, and execute multi-step tasks
• Implementing tool-calling and function-calling patterns with LLMs
• Creating agent memory systems (conversation history, long-term memory, knowledge bases)
• Orchestrating multi-agent systems or agent handoffs
• Experience with agent evaluation, debugging, and observability