DS ML NLP, LLMs - Engineer / Researcher

На данный момент DS ML NLP, LLMs - Engineer / Researcher / Consultant.

Мои компетенции и опыт

Machine Learning Engineer and Researcher:

- 3 years of experience as a NLP and LLMs(Large Language Models) ML Engineer and Researcher.
- 7 years of experience as an AI and Data Science business consultant.

Master of Science. Modern State of Artificial Intelligence specialization,
Applied Mathematics and Computer Science Department, MIPT State University (МФТИ)

HANDS-ON experience with:

- Self-RAG Agents, AI RAG Bots, and AI Bots, with LangGraph, LangChain, asynchronous code / asyncio, REST API/FastAPI, Streamlit WebUI, Ollama, Docker, and Vector Databases, such as ChromaDB, with various closed domain LLMs API and open-source LLMs.
- Training and fine-tuning of LLMs, including SFT, PEFT (LoRA, QLoRA, Prefix and Prompt tuning)
- GPT 3 - 4o, LLaMA 2 - 3.1, Claude, Gemma, Guanaco, BLOOM, BERT, T5, mT5 (base, instruct, flan models), etc.
- Practice with prompt techniques such as in-context learning (one-shot, zero-shot, few-shot), chain of thought, etc.
- Quantization, Distillation, Pruning LLMs.
- Huggingface, ONNX.
- Natural Language processing (NLP) tasks: Sentiment analysis, Summarization and Dialog summarization, Name Entity Recognition (NER), Question answering (Q&A), Entailment generation and Entailment classification, etc.

• I am familiar with: 
- MLOps tools and inference frameworks
- Aligning language models with such technique as RLHF and DPO, IPO, KTO.
- The current state, reviews, and individual Multimodal LLMs (ViLT, BEiT-3, GPT-4V) and their fine-tuning, especially on Video Q&A tasks.

• Research experience:
- Research and testing LLMs on the learning and reasoning abilities at the inference stage.
- Authored two scholarly articles on LLMs for AI conferences.
- Identified and addressed a significant issue in LLMs testing methods.

ARTICLES & PUBLICATIONS
• ”Whether Large Language Models Learn at the Inference Stage?”, September 2023
• ”LLMs. Learning and Reasoning at the Inference Stage”, June 2023 
• ”A Brief Overview of Few-Shot Prompting in Large Language Models”, May 2023