Prediktive -Latin America, LATAM, United StatesWe are looking for a Machine Learning Engineer, based in Latin America to work on a long-term project for one of our clients, a real estate marketplace based in Los Angeles, California.
Our client serves as a resource for individuals aspiring to establish a long-term financial foundation and directly support the communities they care about most.
Responsibilities Play a critical role in integrating LLM models to parse unstructured information from financial documents.Build scalable pipelines for parsing and processing large volumes of documents.Build APIs to expose parsed financial data for downstream applications.Requirements Advanced Level of English4+ years of hands-on experience with models like GPT, BERT, or similar LLMs.Knowledge of strategies to manage LLM input/output length limitations.Experience parsing unstructured data, named entity recognition (NER), text classification, and sequence tagging.Experience using tokenization, embeddings, and transformers.Good working knowledge of techniques for cleaning and structuring raw data, including extracting structured data (e.g., tables, numbers) from unstructured documents, and resolving challenges with OCR and noisy data.Proficiency in Python and ML libraries (e.g., Hugging Face, PyTorch, TensorFlow).Bonus Points Bachelor's Degree in Computer Science, Systems Engineering or related fields.Experience working with PDFs and extracting content.Familiarity integrating ML techniques beyond LLMs - rule-based systems for edge cases and custom models for specific parsing tasks when LLMs are insufficient.Familiarity with tools and frameworks such as: NLP (Hugging Face, SpaCy, NLTK), OCR (Tesseract, AWS Textract, Google Vision API), Data Structuring (Pandas, PyPDF2, Tabula, Camelot for table extraction), LLM APIs (OpenAI API, Azure Cognitive Services, Anthropic's Claude).What we offer Long term positionsCompensation in USDPaid time offCool clients and productsWork with great engineers
#J-18808-Ljbffr