SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
AlphaSense is seeking a Senior Software Engineer to join the Content Engineering team, which specializes in large-scale content processing using cutting-edge AI technologies. The role focuses on leveraging NLP and Large Language Models (LLMs) for document understanding tasks including classification, text categorization, data extraction, and semantic similarity analysis.
You will deploy production-grade LLM-based solutions, fine-tune state-of-the-art models for specialized data extraction from confidential business documents, and develop innovative prompt engineering strategies to maximize reliability and performance. Key responsibilities include solving engineering challenges inherent to LLM solutions such as latency optimization and handling non-deterministic outputs, implementing and scaling data processing pipelines to handle high volumes of complex documents with accuracy and efficiency, and integrating LLM observability tools to trace, debug, and monitor extractions.
You will own systems end-to-end, including design, coding, monitoring, testing, CI/CD, and on-call support. The ideal candidate has 4+ years of software engineering experience primarily with Python, 2+ years of hands-on experience building and deploying LLM applications, and deep expertise in prompt engineering and structured JSON extraction from text. Proficiency with classic NLP libraries (NER, text classification, preprocessing) and experience with LLM tracing platforms like Langfuse are required.
Desirable qualifications include experience with async data processing at scale, frameworks like LangChain and HuggingFace, backend web frameworks such as FastAPI, cloud infrastructure, microservices architecture, Docker, Kubernetes, and domain-specific data extraction experience (financial reports, news). An advanced degree in Computer Science, Computer Engineering, or AI/ML is preferred, though equivalent industry experience is considered.
