Marcus M.

Key Skills

Software

Don't disclose

Other

Top Subject Matter

LLM workflows

structured outputs

prompt engineering

Top Data Types

Text

Top Task Types

Text Generation

Entity (NER) Classification

Classification

Freelancer Overview

Founder | AI Product Builder — Moody Creative Technologies. Brings 7+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose and Other. Education includes Master of Science, National University (2022) and N/A, Le Wagon (2025). AI-training focus includes data types such as Text, Computer Code, and Programming and labeling workflows including Text Generation, Entity (NER) Classification, and Evaluation.

Labeling Experience

Founder and AI Product Builder - Moody Creative Technologies

TextClassification

Built LLM workflow tools and AI-enabled products using Python and FastAPI. Designed structured outputs and prompt logic to improve reliability and guard against failure modes. Collaborated on rapid prototyping focused on speed, clarity, and measurable model performance. • Developed LLM workflow components and API services with Python and FastAPI • Implemented structured output patterns and guardrails for more consistent results • Conducted model evaluation and prompt workflow iteration for operational readiness • Prototyped AI features for real-world usage and faster execution

2025 - Present

Developer | Streamlit Project — Echologic Online Logic Game

Other

Designed and implemented a Python/Streamlit logic game intended to support memory and reasoning tasks. Created a scalable puzzle system spanning multiple difficulty levels, implying controlled generation and validation of task outputs. Developed a clean interface for web and mobile play using a reasoning-oriented application. • Implemented puzzle mechanics across five difficulty levels • Structured game logic to support reasoning and memory • Built an interactive UI using Streamlit • Created reusable code patterns for evaluation of user performance

2025 - 2025

AI Developer | NLP Project — Vibe+ Personality Classifier

Don't discloseTextEntity (NER) Classification

Built an NLP text classifier using a dataset derived from social media tweets. Trained and evaluated a personality classifier model to improve predictive accuracy. Reported substantial accuracy gains, indicating iterative model tuning and performance assessment. • Constructed a classifier pipeline from 7,500 tweets • Tuned the model to improve accuracy by ~70% • Used AI methods to support clearer interpretation • Performed model evaluation to guide improvements

2025 - 2025

Founder | AI Product Builder — Moody Creative Technologies

Don't discloseTextText Generation

Built LLM workflow tools in Python and FastAPI, focusing on structured outputs and prompt logic. Implemented guardrails to improve reliability of AI responses and enable consistent structured results. Worked on prompt workflows and system behaviors rather than traditional manual dataset annotation. • Developed LLM workflow functionality and evaluation-ready outputs • Designed structured output formats for clearer AI results • Added guardrails and prompt logic to reduce failure modes • Supported AI operations and prompt workflow execution

2020 - 2022

Education

Work History