Machine Learning Engineer

AP Executive are working with an institutional-grade crypto intelligence terminal. They aim to provide comprehensive news coverage, research, and data products to industry participants. We are seeking a machine learning engineer with deep expertise in building production AI systems that handle complex, real-time data at scale.

The Role

You will architect and build an AI system for cryptocurrency market intelligence - a sophisticated application that synthesizes multiple data sources including market data, news, and technical analysis to answer complex market questions that professional traders ask.

You will play a key role in defining the technical approach for this project - helping evaluate whether to leverage existing foundation models with fine-tuning, build custom architectures, or use API-based solutions. This is a greenfield opportunity where you'll contribute significantly to architectural decisions based on accuracy requirements, cost constraints, latency targets, and scalability needs.

Key responsibilities include:

Contributing to defining the overall technical approach and architecture - helping evaluate whether to use existing foundation models (GPT-4, Claude, Gemini) with RAG, fine-tune open-source models, or build custom solutions from scratch
Designing and implementing a RAG system that retrieves and synthesizes information from multiple heterogeneous data sources in real-time
Building strict hallucination prevention mechanisms - the system must only make claims backed by source data and explicitly state uncertainty when appropriate
Implementing sophisticated time-awareness and data freshness tracking across all queries and responses
Creating systems that synthesize multi-dimensional market signals into coherent analysis
Integrating outputs from quantitative systems and presenting them effectively through conversational interfaces
Building scenario-based reasoning capabilities that provide probabilistic assessments rather than deterministic predictions
Collaborating with data scientists and quantitative researchers to define optimal data structures and integration patterns
Creating conversational flows that intelligently gather context, ask clarifying questions, and guide users to relevant insights
Developing prompt engineering frameworks and guardrails that maintain consistent, reliable behavior
Building comprehensive monitoring systems to detect degraded performance, factual errors, or hallucinations in production
Implementing source citation and data provenance tracking throughout the system
Designing APIs and interfaces between the AI system and other infrastructure components

Autonomy and Approach

You will have significant input in defining the technical approach to this project. You'll help evaluate whether to use existing foundation models via APIs, fine-tune open-source models, or build custom architectures based on accuracy requirements, cost constraints, latency targets, and scalability needs.

Beyond architecture, you'll contribute to choosing RAG frameworks, retrieval strategies, and implementation patterns. However, certain principles are non-negotiable: factual accuracy above all else, explicit uncertainty handling, source citations for claims, and scenario-based reasoning rather than overconfident predictions. These aren't optional features - they're fundamental requirements for building systems that users trust with high-stakes decisions.

You will help make critical decisions about when to use retrieval versus when models can reason directly, how to structure prompts for consistent behavior, and how to balance response quality with latency.

Candidate Profile

Qualifications Experience and Track Record:

Minimum 5+ years of experience building and deploying machine learning systems in production
Proven experience contributing to architectural decisions for AI/ML projects - helping evaluate API-based solutions, fine-tuning approaches, and custom implementations based on product requirements
Demonstrated track record of evaluating and selecting appropriate LLM solutions (foundation model APIs, open-source models, custom architectures) for production use cases
Proven experience with large language models, RAG architectures, or conversational AI systems in production
Track record of building AI products that prioritized factual accuracy and appropriate handling of uncertainty
Experience working with real-time data integration and multi-source information synthesis
Portfolio of implemented LLM systems or contributions demonstrating technical depth in modern AI architectures

LLM and RAG Expertise:

Strong proficiency with modern LLM frameworks (LangChain, LlamaIndex, or custom implementations)
Deep understanding of RAG architectures including retrieval strategies, re-ranking, and context window management
Experience with vector databases and semantic search (Pinecone, Weaviate, Chroma, or similar)
Practical knowledge of prompt engineering, few-shot learning, and techniques for consistent LLM behavior
Understanding of hallucination detection and mitigation strategies in production systems
Experience with LLM fine-tuning, RLHF, or model adaptation techniques (highly valued)
Knowledge of LLM inference optimization and deployment (vLLM, TensorRT-LLM, or similar)

Technical Capabilities:

Strong proficiency in Python and modern ML frameworks (PyTorch, TensorFlow, or JAX)
Experience building conversational flows and multi-turn dialogue systems
Solid understanding of natural language processing fundamentals
Proficiency with API design for AI systems including streaming responses and session management
Experience with real-time data pipelines and event-driven architectures
Knowledge of caching strategies and performance optimization for LLM systems
Understanding of evaluation frameworks for generative AI (factual accuracy, relevance, hallucination detection)
Familiarity with MLOps practices specific to LLM deployments

Data and Domain Knowledge:

Experience working with time-series data and understanding temporal context in queries
Familiarity with financial markets, trading concepts, or similar high-stakes decision-making domains
Ability to understand and communicate quantitative signals and technical metrics in user-friendly ways
Experience integrating with analytical systems and understanding their outputs (statistical measures, risk metrics, validation results)
Understanding of how to structure and retrieve information from heterogeneous data sources
Experience building systems that cite sources and maintain data provenance
Willingness to develop deep domain expertise quickly

Engineering Excellence:

Ability to write production-quality code with appropriate testing and error handling
Experience with monitoring and observability for AI systems (tracking hallucinations, quality degradation, etc.)
Strong debugging skills for complex AI systems where failures are often subtle
Capacity to design systems that maintain quality while scaling to many concurrent users
Understanding of security considerations for AI systems (prompt injection, data leakage, etc.)
Experience building APIs and integration layers between different systems (AI models, databases, quant engines)
Ability to work cross-functionally with specialized teams (quants, data scientists, researchers) and integrate their outputs seamlessly

Product and Research Capabilities:

Ability to evaluate and help select appropriate technical approaches based on product requirements, cost constraints, and performance targets
Experience contributing to build-vs-buy decisions for ML components and helping justify architectural choices
Ability to translate user needs into technical requirements for AI systems
Experience designing conversational experiences that guide users effectively
Skill in evaluating subjective AI outputs and establishing quality metrics
Understanding of when to build custom solutions versus when to use existing LLM APIs
Capacity to stay current with rapidly evolving LLM research and apply relevant techniques

To apply for this position please send your CV to Don Fletcher via