Celal Turkmen

Data Scientist at Jotform Inc.

Built AI agents as conversational and workflow automation tools to streamline business tasks, including automated form creation and filling. Contributed to the Teach Your Agent feature, enabling AI agents to detect user instructions during conversations and autonomously update knowledge base or action logic. Resolved 50+ production issues through debugging, testing, and log analysis using Kibana. Wrote and maintained 5,000+ lines of production-level code in Python and PHP. Designed and implemented automated tests to evaluate and improve prompt performance. Developed parallel execution pipelines for LLMs, cutting response latency by 50%.

Data Scientist at Jotform Inc.

Built AI agents as conversational and workflow automation tools to streamline business tasks, including automated form creation and filling. Contributed to the Teach Your Agent feature, enabling AI agents to detect user instructions during conversations and autonomously update their knowledge base or action logic. Resolved 50+ production issues through debugging, testing, and log analysis using Kibana. Wrote and maintained 5,000+ lines of production‑level code in Python and PHP. Designed and implemented a suite of automated tests to evaluate and improve prompt performance. Developed parallel execution pipelines for LLMs, cutting response latency by 50%.

Data Scientist at Jotform Inc.

Built AI agents as conversational and workflow automation tools to streamline business tasks, including automated form creation and filling. Contributed to the Teach Your Agent feature, enabling AI agents to detect user instructions during conversations and autonomously update their knowledge base or action logic. Resolved 50+ production issues through debugging, testing, and log analysis using Kibana. Wrote and maintained 5,000+ lines of production-level code in Python and PHP. Designed and implemented a suite of automated tests to evaluate and improve prompt performance. Developed parallel execution pipelines for LLMs, cutting response latency by 50%.

Data Scientist at Jotform Inc.

Built AI agents as conversational and workflow automation tools to streamline business tasks, including automated form creation and filling. Contributed to the Teach Your Agent feature, enabling AI agents to detect user instructions during conversations and autonomously update knowledge bases or action logic. Resolved 50+ production issues through debugging, testing, and log analysis with Kibana. Wrote and maintained 5,000+ lines of production-level Python and PHP code. Designed and implemented automated tests to evaluate and improve prompt performance. Developed parallel execution pipelines for LLMs, reducing response latency by 50%.

Project Lead at yolla.tech

Developed a parcel service search engine to identify optimal options based on cost, speed, and sustainability within a team of five; technologies included React, FastAPI, PostgreSQL, Docker, and AWS EC2 to automate and significantly reduce the exhaustive parcel service search process.

Machine Learning Intern at DataBoss Analytics & Security

Fine-tuned a Retrieval-Augmented Generation (RAG) application powered by local LLMs integrated with a Chroma vector database, achieving a 60% reduction in information retrieval time. Developed localized solutions for the company’s custom chatbot, ensuring compliance with data privacy policies. Gained experience in LLMs, LangChain, Docker, and Python, writing 1,000+ lines of code.

Machine Learning Intern at DataBoss Analytics & Security

Fine-tuned a Retrieval-Augmented Generation (RAG) application powered by local LLMs integrated with a Chroma vector database, achieving a 60% reduction in information retrieval time. Developed localized solutions for the company’s custom chatbot, ensuring compliance with the data privacy policies. Gained experience in LLMs, LangChain, Docker, and Python, writing 1,000+ lines of code.

Machine Learning Intern at DataBoss Analytics & Security

Fine-tuned a Retrieval-Augmented Generation (RAG) application powered by local LLMs integrated with a Chroma vector database, achieving a 60% reduction in information retrieval time. Developed localized solutions for the company’s custom chatbot, ensuring compliance with data privacy policies. Gained experience in LLMs, LangChain, Docker, and Python, writing 1,000+ lines of code.

Machine Learning Intern at DataBoss Analytics & Security

Fine-tuned a Retrieval-Augmented Generation (RAG) application powered by local LLMs integrated with a Chroma vector database, achieving a 60% reduction in information retrieval time. Developed localized solutions for the company’s custom chatbot, ensuring compliance with data privacy policies. Gained experience in LLMs, LangChain, Docker, and Python, writing 1,000+ lines of code.

Team Lead / Project Lead at BilConnect

Led the development team in creating a social website for university students; used ASP.NET and C#; gained experience in requirements elicitation, MVC, design patterns, UML diagrams, and object modeling. Created CI/CD pipelines and implemented GitHub Workflows for automated testing and streamlined deployment.

Research Intern at National Taiwan University

Trained semantic segmentation models using PyTorch, U-Net architecture, and MONAI library, processing 100+ CT volumes for computer vision and deep learning applications in medical imaging. Worked with a team of 5 researchers to improve model accuracy by 85%, reducing prediction error rates by 55%. Awarded TEEP scholarship from the Taiwan Ministry of Education.

Research Intern at National Taiwan University

Trained semantic segmentation models using PyTorch, U-Net architecture, and the MONAI library, processing 100+ CT volumes for research on computer vision and deep learning applications in medical imaging. Worked with a team of 5 researchers to improve model accuracy by 85%, reducing prediction error rates by 55%. Awarded TEEP (Taiwan Experience Education Program) scholarship from the Taiwan Ministry of Education.

Research Intern at National Taiwan University

Trained semantic segmentation models using PyTorch, U-Net, and MONAI; processed 100+ CT volumes for medical imaging research. Worked with a team of 5 researchers to improve model accuracy by 85% and reduce prediction error by 55%. Awarded the TEEP scholarship from the Taiwan Ministry of Education.