Josef Pícha
Machine Learning | NLP | LLM | Deep Learning Engineer
Prague,10
Summary
Dynamic and results-driven Machine Learning Engineer with strong professional experience specializing in advanced Retrieval-Augmented Generation (RAG) systems, large language models (LLMs), Generative AI and machine learning algorithms. With a Master’s in Neural Networks and Machine Learning from the University of Birmingham, brings deep expertise in state-of-the-art machine learning techniques, natural language processing, and generative AI technologies. Technical acumen, combined with hands-on experience in real-world projects, allows to design scalable, impactful AI solutions that enhance productivity and streamline processes.
Overview
4
4
years of professional experience
5
5
years of post-secondary education
3
3
Certifications
2
2
Languages
Work History
Chief Technology Officer | Generative AI Engineer
Scentia
, Sweden
06.2024 - Current
- Spearheaded the development of an AI-powered school platform to enhance teacher-student collaboration, utilizing generative LLM technology to transform educational experiences.
- Designed and implemented advanced Retrieval-Augmented Generation (RAG) systems leveraging cutting-edge tools like LangChain, LangSmith, Azure Functions, and Python frameworks (FastAPI, Flask).
- Directed a team of three developers, fostering a collaborative environment to deliver innovative and scalable AI-driven solutions.
- Engineered and maintained the platform's core AI functionalities, ensuring high performance, reliability, and alignment with user needs.
- Established and managed cloud-based infrastructure on Microsoft Azure, optimizing cost efficiency and operational effectiveness.
- Defined and executed the company’s technological roadmap, aligning product development with long-term strategic goals.
- Championed agile methodologies to streamline workflows, enhance team productivity, and improve development cycles.
- Conducted rigorous testing and validation of AI models to ensure compliance with quality standards and deliver superior user experiences.
- Collaborated with cross-functional teams to identify emerging trends and integrate state-of-the-art technologies, maintaining a competitive edge in the edtech market.
- Cultivated a culture of innovation within the organization, mentoring junior developers and fostering growth in AI expertise.
AI Architect & Engineer
Cetin
Prague, Hlavni mesto Praha
03.2024 - Current
- Designed and developed end-to-end machine learning (ML) solutions tailored to diverse use cases across the organization, starting from problem scoping to final deployment.
- Collaborated with clients to understand their challenges and envisioned solutions, creating bespoke ML architectures based on provided data and objectives.
- Delivered four successful AI-driven products, including:
- Anomaly Detection System: Monitored over 27,000 unique data streams using advanced statistical methods and predictive models like Prophet and Isolation Forest.
- HR Chatbot: Implemented a medium-scale Retrieval-Augmented Generation (RAG) solution to enhance HR processes.
- AI Testing Assistant: Developed an advanced RAG-based solution incorporating image recognition and few-shot learning to optimize end-to-end testing workflows.
- Laws Chatbot: Led the development of a large-scale chatbot for legal applications, currently in the early stages of implementation. - Defined scalable and efficient ML pipelines, integrating advanced statistical models and innovative techniques to meet complex client requirements.
- Optimized workflows by architecting robust ML solutions, enabling seamless integration with existing infrastructure and tools.
- Partnered with cross-functional teams, ensuring a deep understanding of client goals and translating them into actionable technical frameworks.
- Applied advanced AI methodologies, including RAG systems and statistical modeling, to drive operational efficiency and deliver measurable business value.
- Mentored junior team members, fostering growth in ML and AI expertise across the organization.
- Maintained a proactive approach to exploring emerging technologies, incorporating best practices to deliver state-of-the-art solutions.
LLM & Machine Learning Software Engineer
Adastra AI (Blindspot.ai)
Prague, Hlavni mesto Praha
09.2023 - Current
- Designed and implemented advanced Retrieval-Augmented Generation (RAG) systems leveraging state-of-the-art LLM models and vector databases on cloud platforms like Azure and AWS, delivering high-performance AI solutions.
- Engineered traditional machine learning models for diverse use cases, including anomaly detection, product demand forecasting, and clustering-based recommendation systems, optimizing decision-making processes.
- Managed version control for machine learning projects using Git, maintaining well-structured and organized codebases to support collaborative development.
- Conducted rigorous code reviews to ensure adherence to best practices and high-quality standards across all project deliverables.
- Created user-friendly interfaces for AI models, improving accessibility and usability for non-technical stakeholders.
- Implemented innovative machine learning solutions for complex datasets, achieving enhanced prediction accuracy and actionable insights.
- Utilized advanced visualization techniques to simplify analysis of highly numeric data, providing clear and intuitive reports for stakeholders.
- Developed scalable and efficient data pipelines for real-time deployment of machine learning models, ensuring consistent performance and reliability.
- Evaluated and optimized machine learning libraries and tools, selecting the best resources for specific project needs to enhance model performance.
- Applied ensemble methods such as bagging and boosting to improve the predictive capabilities of traditional machine learning models.
- Explored and incorporated emerging trends in AI and ML, including advancements in LLMs, into ongoing projects to maintain cutting-edge solutions.
- Collaborated with cross-functional teams to design and deploy robust AI systems, ensuring seamless integration with existing infrastructure and workflows.
- Conducted thorough data analysis to identify patterns and trends, driving the development of effective machine learning models tailored to business needs.
Deep Learning Thesis (Master of Science)
University of Birmingham
Birmingham, United Kingdom
04.2023 - 09.2023
- Conducted advanced research on state-of-the-art Question Answering (QA) models, focusing on their performance with specialized medical datasets.
- Evaluated the capabilities of cutting-edge natural language processing (NLP) models, including ELECTRA, XLNet, and T5, to retrieve and process complex medical information efficiently and accurately.
- Designed and implemented robust experiments to analyze the scalability, accuracy, and precision of QA systems in addressing the unique challenges of medical terminology, symptoms, and diagnostic processes.
- Developed and preprocessed specialized medical datasets to ensure data quality and alignment with the specific requirements of QA systems.
- Compared and benchmarked performance metrics across various models, highlighting their potential for real-world medical applications.
- Investigated the complexities of medical data, proposing innovative solutions to improve information retrieval and response generation in healthcare contexts.
- Explored the transformative potential of QA systems in revolutionizing medical diagnosis, research, and literature synthesis.
- Authored a comprehensive thesis detailing methodologies, findings, and recommendations for future advancements in medical QA technologies.
- Applied a multidisciplinary approach, integrating machine learning, NLP, and domain-specific expertise in medicine, to address complex challenges in information retrieval.
- Contributed to the academic body of knowledge on QA systems, emphasizing the importance of precision and efficiency in medical applications.
Software Engineer
Vocalls
05.2022 - 04.2023
- Developed scalable and maintainable code, ensuring long-term stability of the software.
- Integrated new technologies into existing systems, increasing capabilities and improving overall performance.
- Developed reusable components that significantly reduced development effort on multiple projects.
- Implemented effective debugging strategies, resulting in fewer software defects and increased reliability.
- Enhanced user experience with intuitive interface design and responsive web applications.
- Improved software performance by identifying and resolving bottlenecks in the code.
- Analyzed proposed technical solutions based on customer requirements.
Software Engineering Project (Bachelor of Science)
University of Portsmouth
01.2022 - 05.2022
- Designed and developed a Discord Quiz Bot using JavaScript, enabling users to participate in interactive quiz games within Discord servers, addressing both educational needs and pandemic-related challenges.
- Conducted a comprehensive literature review to analyze similar systems, gaining insights into existing Discord bot technologies and identifying key areas for innovation.
- Defined and documented system requirements through questionnaires, analysis of existing tools, and elicitation techniques to ensure the bot met user needs and technical standards.
- Architected the system's design, including GUI mockups and backend frameworks, ensuring a seamless user experience and robust functionality.
- Selected and implemented an appropriate development methodology, justifying its application and ensuring structured and efficient project execution.
- Developed and tested the Discord bot iteratively, documenting progress and resolving technical challenges to achieve the project’s functional and performance goals.
- Evaluated the bot against defined requirements, ensuring compliance with Discord bot technologies and validating its effectiveness as an educational tool.
- Produced a detailed report covering the project’s methodologies, design, development process, testing, and future improvement opportunities.
- Highlighted the bot’s potential applications in education and engagement, demonstrating its utility beyond the initial pandemic-focused context.
Self Employed Software Engineer
Self-employeed
11.2021 - 01.2022
- Developed scalable and maintainable code, ensuring long-term stability of the software.
- Integrated new technologies into existing systems, increasing capabilities and improving overall performance.
- Developed reusable components that significantly reduced development effort on multiple projects.
- Implemented effective debugging strategies, resulting in fewer software defects and increased reliability.
Software Engineering Intern
Argo22
06.2021 - 11.2021
- Participated in agile development processes, effectively adapting to changing requirements while maintaining high-quality results.
- Documented development procedures, creating valuable reference materials for future projects or team members joining the team.
- Collaborated with software engineers to develop and test application procedures for system efficiency.
- Enhanced software performance by optimizing algorithms and streamlining code.
Education
Master of Science - Neural Networks & Machine Learning
University of Birmingham
Birmingham 09.2022 - 09.2024
Bachelor of Science - Software Engineering
University of Portsmouth
Portsmouth 09.2019 - 06.2022
Skills
Technology roadmapping
Software development
Machine learning
System architecture
Python
RAG
LLM
Microsoft Azure
GEN AI
Vector Search
Embedding Models
Deep learning algorithms
Machine learning frameworks
SQL databases
Scikit-learn
Data preprocessing
Git version control
Model evaluation
Neural networks
Feature engineering
Natural language processing
Machine learning integration
Time series analysis
Data science principles implementation
Data visualization
Agile methodologies
Cloud computing
Pytorch
Pandas
Object-oriented programming
Javascript
Software development life cycle (SDLC)
JIRA
API development
Testing and debugging
Amazon web services
Version control systems expertise
Experience with deployment automation
Databases: MongoDB, Postresql, Cosmos DB
Software deployment
Generative adversarial networks
Supervised learning
Unsupervised learning
NoSQL databases
Hyperparameter tuning
Data analytics
Predictive modeling
Multitasking Abilities
Databricks
Langchain
Langsmith
LangGraph
Docker
Certification
Udemy Deep Learning Bootcamp Certificate
Timeline
Udemy Deep Learning Bootcamp Certificate
12-2024
Udemy Pytorch for Deep Learning Certificate
12-2024
Generative AI in Databricks
11-2024
Chief Technology Officer | Generative AI Engineer
Scentia
06.2024 - Current
AI Architect & Engineer
Cetin
03.2024 - Current
LLM & Machine Learning Software Engineer
Adastra AI (Blindspot.ai)
09.2023 - Current
Deep Learning Thesis (Master of Science)
University of Birmingham
04.2023 - 09.2023
Master of Science - Neural Networks & Machine Learning
University of Birmingham
09.2022 - 09.2024
Software Engineer
Vocalls
05.2022 - 04.2023
Software Engineering Project (Bachelor of Science)
University of Portsmouth
01.2022 - 05.2022
Self Employed Software Engineer
Self-employeed
11.2021 - 01.2022
Software Engineering Intern
Argo22
06.2021 - 11.2021
Bachelor of Science - Software Engineering
University of Portsmouth
09.2019 - 06.2022