For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

LoginCreate account
Invite to Job
M

Mike S.

AI Research Engineer

Germany flagHeidelberg, Germany
$36.00/hrIntermediateSuperannotate

Key Skills

Software

SuperAnnotateSuperAnnotate

Top Subject Matter

Large Language Models
RLHF / AI Alignment
Multi-Agent Systems

Top Data Types

TextText
DocumentDocument

Top Task Types

RLHF
Fine Tuning
Evaluation Rating

Freelancer Overview

AI Research Engineer (Internship→Consultant). Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal and Proprietary Tooling. Education includes Master of Science, Georg-August-Universität Göttingen (2026) and Bachelor of Science, Conservatoire National des Arts et Métiers (CNAM) (2021). AI-training focus includes data types such as Text and labeling workflows including RLHF and Prompt + Response Writing (SFT).

IntermediateEnglishCantoneseChinese Mandarin

Labeling Experience

Annotation Reviewer

TextRLHF
RLHF annotation and review for a large language model alignment project. Promoted from Annotator to Reviewer within 2 weeks based on annotation quality and throughput. Scope: Evaluate pairs of LLM-generated responses against user prompts using a structured rubric covering instruction following, completeness, factual accuracy, coherence, relevance, efficiency, articulation, and length appropriateness. Each task involves writing detailed Strengths and Areas of Improvement with cited excerpts, assigning 1-5 quality scores per response, and determining preference rankings with justification. As Reviewer: Independently complete full annotations, then audit other annotators' work — approving, providing corrective feedback, or flagging substantial misses in Strengths/AOIs assessments. Quality measures: Fact-checking all factual claims against primary sources, cross-referencing annotations against a 15-section internal style guide, and running systematic QC checklists before submission. Bilingual coverage (English and Simplified Chinese).

RLHF annotation and review for a large language model alignment project. Promoted from Annotator to Reviewer within 2 weeks based on annotation quality and throughput. Scope: Evaluate pairs of LLM-generated responses against user prompts using a structured rubric covering instruction following, completeness, factual accuracy, coherence, relevance, efficiency, articulation, and length appropriateness. Each task involves writing detailed Strengths and Areas of Improvement with cited excerpts, assigning 1-5 quality scores per response, and determining preference rankings with justification. As Reviewer: Independently complete full annotations, then audit other annotators' work — approving, providing corrective feedback, or flagging substantial misses in Strengths/AOIs assessments. Quality measures: Fact-checking all factual claims against primary sources, cross-referencing annotations against a 15-section internal style guide, and running systematic QC checklists before submission. Bilingual coverage (English and Simplified Chinese).

2026 - Present

Teaching Assistant (Applied NLP)

TextPrompt Response Writing SFT
I designed and delivered Jupyter tutorials addressing embeddings, attention mechanisms, LoRA, SFT, DPO, and transformer interpretability. As a teaching assistant, I contributed code and exercises for students to participate in LLM-from-scratch projects and hands-on data annotation tasks. The tutorials focused on hands-on text generation, prompt writing, and model understanding for over 100 MSc students. • Wrote and labeled prompts/responses for transformer model assignments • Generated instructional datasets for student exercises • Assisted in designing and reviewing LLM model outputs for learning libraries • Supported classroom annotation activities on NLP models

I designed and delivered Jupyter tutorials addressing embeddings, attention mechanisms, LoRA, SFT, DPO, and transformer interpretability. As a teaching assistant, I contributed code and exercises for students to participate in LLM-from-scratch projects and hands-on data annotation tasks. The tutorials focused on hands-on text generation, prompt writing, and model understanding for over 100 MSc students. • Wrote and labeled prompts/responses for transformer model assignments • Generated instructional datasets for student exercises • Assisted in designing and reviewing LLM model outputs for learning libraries • Supported classroom annotation activities on NLP models

2024 - 2025

Graduate Researcher & Core Developer

TextPrompt Response Writing SFT
As a core developer for an academic Agentic RAG AI Tutor, I implemented pipelines that transformed course materials into enriched text datasets for LLM training. I generated and labeled prompts and responses for tutoring, facilitating improved question answering and proactive assistance modes. The work included constructing Markdown knowledge graphs and contributing to LLM-driven instructional content creation. • Created and annotated textual data for educational AI tutoring • Built automated enrichment pipelines converting source documents to training data • Produced labeled knowledge graphs for use in tutoring agents • Supported text generation and QA model fine-tuning for academic research

As a core developer for an academic Agentic RAG AI Tutor, I implemented pipelines that transformed course materials into enriched text datasets for LLM training. I generated and labeled prompts and responses for tutoring, facilitating improved question answering and proactive assistance modes. The work included constructing Markdown knowledge graphs and contributing to LLM-driven instructional content creation. • Created and annotated textual data for educational AI tutoring • Built automated enrichment pipelines converting source documents to training data • Produced labeled knowledge graphs for use in tutoring agents • Supported text generation and QA model fine-tuning for academic research

2024 - 2025

AI Research Engineer (Internship→Consultant)

TextRLHF
I built an end-to-end RLHF pipeline on an 8-GPU AWS cluster involving SFT, GRPO/PPO, evaluation, and OpenAI-compatible serving. I fine-tuned large language models using carefully curated mixed datasets and implemented evaluation frameworks to score generated outputs. The process led to precise model performance improvements and production-quality data pipelines. • Developed RLHF workflow using reinforcement and supervised fine-tuning stages • Curated and labeled reasoning-based text datasets for LLM training • Implemented LLM-as-judge evaluation using benchmarks like WritingBench and MT-Bench • Contributed to third-party audit pipeline labeling with smart contract taxonomy and report annotation

I built an end-to-end RLHF pipeline on an 8-GPU AWS cluster involving SFT, GRPO/PPO, evaluation, and OpenAI-compatible serving. I fine-tuned large language models using carefully curated mixed datasets and implemented evaluation frameworks to score generated outputs. The process led to precise model performance improvements and production-quality data pipelines. • Developed RLHF workflow using reinforcement and supervised fine-tuning stages • Curated and labeled reasoning-based text datasets for LLM training • Implemented LLM-as-judge evaluation using benchmarks like WritingBench and MT-Bench • Contributed to third-party audit pipeline labeling with smart contract taxonomy and report annotation

2024 - 2024

Education

G

Georg-August-Universität Göttingen

Master of Science, Applied Computer Science (Data Science)

Master of Science
2022 - 2026
C

Conservatoire National des Arts et Métiers (CNAM)

Bachelor of Science, Informatics and Software Engineering

Bachelor of Science
2017 - 2021

Work History

N

NEC Laboratories Europe

Master's Thesis Researcher

Heidelberg
2025 - 2026
L

Lufthansa Technik

Data Engineer Intern

Hamburg
2025 - 2025