• Adapt and fine-tune custom LLMs (e.g., GPT-4, LLAMA-70b) for specific tasks like document analysis and data extraction, using specialized frameworks for LLM development (Langchain, Meta-Instruct)
• Implement prompt engineering strategies to enhance model performance and accuracy
• Design and implement custom preprocessing modules to prepare large documents (e.g., PDFs) for analysis
• Deploy and manage ML/LLM models and applications on cloud services such as AWS and Azure, while optimizing infrastructure for scalability, reliability, and cost-efficiency
• Implement classical NLP methods: NER, Word2Vec, Vectorization, LSTM/RNN models for classification/vectorization and others

• Proficiency in Python with extensive experience in PyTorch, numpy, pandas
• Hands-on experience with CUDA is a plus
• Expertise in data preprocessing techniques for NLP tasks, using tools like NLTK, spaCy, Langchain, Huggingface
• Experience with NLP tasks such as tokenization, NER, part-of-speech tagging, sentiment analysis, and embeddings like Word2Vec, GloVe, BERT
• Hands-on experience with GPT models, including fine-tuning and deployment, strong theoretical knowledge of Transformers and LLMs
• Proficiency in LangChain for building and deploying LLM applications and creating complex pipelines
• Understanding of model evaluation metrics, tuning techniques, API development, and integration
• Familiarity with RESTful APIs, cloud services like Azure, AWS, and proficiency with Git and collaborative platforms like GitHub or GitLab

At Vention, we assemble senior-level, dedicated teams of developers to help fast-growing startups and innovative enterprises drive impact and achieve their goals. We’ve delivered solutions across multiple domains, including FinTech, PropTech, AdTech, HealthTech, e-commerce, and more.

Our Data team works with clients to create data platforms from scratch or modify and update existing platforms. The tech stack depends on the project, but we mainly use Spark (along with Scala, Python, or Java) – as well as Apache Kafka, Apache Cassandra, Apache Hadoop, Apache Parquet, and AWS.

Internal knowledge transfer activities are conducted within the Data Engineering Family (which includes data practice & data competency) – it is a space for all of our specialists to share their experiences, learn new skills, host meetups, mentor others, and more.