Requirements

Expected technologies

Python

TensorFlow

NLP

LLM

Azure DevOps

LangChain

RAG

Optional technologies

Deep Learning

Computer Vision

Our requirements

• Minimum 6+ years of hands-on experience in data science.
• Strong programming skills in Python, PyTorch, TensorFlow, and related libraries.
• Solid understanding of object-oriented design patterns, concurrency/multithreading, and scalable AI/GenAI model deployment.
• Proficiency in NLP techniques: RegEx, SpaCy, NLTK, text representation, and semantic extraction.
• Hands-on experience developing, training, and fine-tuning LLMs and AI models.
• Practical knowledge of CNN, RNN, GANs, RAG, LangChain, and Transformers.
• Expertise in prompt engineering techniques and vector databases.
• Familiarity with Azure Cloud Computing Platform.
• Experience with Docker, Kubernetes, and CI/CD pipelines.
• Experience with Deep Learning, Computer Vision, CNN, RNN, LSTM.
• Knowledge of vector databases (Milvus, Postgres, etc.) and other database technologies.

Your responsibilities

• Collaborate with engineers, data scientists, and business analysts to understand requirements, refine models, and integrate LLMs into AI solutions.
• Apply Reinforcement Learning with Human Feedback (RLHF) and advanced AI techniques for tax-specific outputs.
• Embed generative AI solutions into consolidation, reconciliation, and reporting processes.
• Leverage LLMs to process and interpret unstructured tax documentation.
• Develop and implement deep learning algorithms for AI solutions.
• Preprocess raw data (text normalization, tokenization, etc.) for NLP model use.
• Setup, train, and fine-tune large language models and other state-of-the-art neural networks.
• Conduct testing, validation, and statistical analysis to ensure accuracy and optimize performance.
• Explore and propose innovative AI use cases to enhance tax functions.
• Partner with tax, finance, IT, and legal teams to integrate AI workflows and comply with regulatory standards.
• Perform model audits to identify and mitigate risks.
• Monitor and optimize generative models for performance and scalability.