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
Mehdi

Mehdi

Computational Biologist AI/ML

USA flagLos Angeles, Usa
$40.00/hrIntermediateOtherRedbrick AISloth

Key Skills

Software

Other
Redbrick AIRedbrick AI
SlothSloth
Internal/Proprietary Tooling

Top Subject Matter

Computational Biologist | LLM Evaluator
Comp Bio + AI/ML | LLM Training & Evaluation
Bio-AI Specialist

Top Data Types

TextText
AudioAudio
ImageImage

Top Task Types

RLHFRLHF
ClassificationClassification
TranscriptionTranscription
Text GenerationText Generation
Fine-tuningFine-tuning
Computer Programming/CodingComputer Programming/Coding
Data CollectionData Collection
Function CallingFunction Calling
Object DetectionObject Detection
Text SummarizationText Summarization
Prompt + Response Writing (SFT)Prompt + Response Writing (SFT)
Evaluation/RatingEvaluation/Rating
Question AnsweringQuestion Answering

Freelancer Overview

Computational Biologist — Rossi. Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Other. Education includes Bachelor of Science, University of California, Los Angeles (2025) and Associate of Science, Irvine Valley College (2022). AI-training focus includes data types such as Text and Audio and labeling workflows including RLHF, Evaluation, and Rating.

IntermediateEnglish

Labeling Experience

Computational Biologist — Rossi

TextRLHF
Trained large language models (LLMs) from scratch across RNN, LSTM, and Transformer architectures using full-cycle data curation and iterative training. Designed and applied reinforcement learning from human feedback (RLHF) and reward modeling to align model behaviors and outputs. Built bespoke domain-specific evaluation benchmarks with prompt engineering and ground-truth scoring for advanced model assessment. • Data lifecycle management encompassing collection, extraction, cleaning, and classification • Implemented structured reasoning and Chain-of-Thought (CoT) analysis with feedback-driven corrections • Developed adversarial evaluation for model robustness and safety • Constructed hand-crafted benchmarks and scoring pipelines using Python and PyTorch

Trained large language models (LLMs) from scratch across RNN, LSTM, and Transformer architectures using full-cycle data curation and iterative training. Designed and applied reinforcement learning from human feedback (RLHF) and reward modeling to align model behaviors and outputs. Built bespoke domain-specific evaluation benchmarks with prompt engineering and ground-truth scoring for advanced model assessment. • Data lifecycle management encompassing collection, extraction, cleaning, and classification • Implemented structured reasoning and Chain-of-Thought (CoT) analysis with feedback-driven corrections • Developed adversarial evaluation for model robustness and safety • Constructed hand-crafted benchmarks and scoring pipelines using Python and PyTorch

2026 - Present

Head of AI Platform, Computational Biology & AI/ML — Xavvi

Text
Researched and evaluated cognitive reasoning patterns, biases, and behavioral functions in large language models using psychological and factual assessment tasks. Developed and executed LLM evaluation infrastructure, including prompt engineering and response validation for model reasoning and factual grounding. Managed Chain-of-Thought debugging and feedback loops to refine LLM outputs and reasoning quality. • Built evaluation pipelines with LangGraph agent orchestration • Conducted bias/behavioral analysis via structured prompt/response tests • Developed and maintained prompt and rating schema for model assessments • Researched cognitive function and debugging strategies using live LLM output analysis

Researched and evaluated cognitive reasoning patterns, biases, and behavioral functions in large language models using psychological and factual assessment tasks. Developed and executed LLM evaluation infrastructure, including prompt engineering and response validation for model reasoning and factual grounding. Managed Chain-of-Thought debugging and feedback loops to refine LLM outputs and reasoning quality. • Built evaluation pipelines with LangGraph agent orchestration • Conducted bias/behavioral analysis via structured prompt/response tests • Developed and maintained prompt and rating schema for model assessments • Researched cognitive function and debugging strategies using live LLM output analysis

2025 - Present

Clinical Trial Evaluation Benchmark (Project Role)

Text
Authored and executed 144 domain-specific prompts for clinical trial analytics, creating a comprehensive evaluation benchmark for LLM performance. Developed scoring pipelines combining LLM-as-judge evaluations with structured schema integration for factual accuracy and granularity. Used hand-crafted ground-truth queries and outputs to accurately measure model reliability and robustness. • Included multi-dimensional scoring: accuracy, completeness, hallucination detection • Implemented answer schema: confidence level, caveats, and source attribution • Built the evaluation platform incorporating CDISC SDTM/ADaM domains, RAG retrieval, and SAS Viya • Used Python, SAS Viya, Pinecone, and OpenAI for analytics and evaluation

Authored and executed 144 domain-specific prompts for clinical trial analytics, creating a comprehensive evaluation benchmark for LLM performance. Developed scoring pipelines combining LLM-as-judge evaluations with structured schema integration for factual accuracy and granularity. Used hand-crafted ground-truth queries and outputs to accurately measure model reliability and robustness. • Included multi-dimensional scoring: accuracy, completeness, hallucination detection • Implemented answer schema: confidence level, caveats, and source attribution • Built the evaluation platform incorporating CDISC SDTM/ADaM domains, RAG retrieval, and SAS Viya • Used Python, SAS Viya, Pinecone, and OpenAI for analytics and evaluation

2024 - 2024

Research Assistant — Irvine Valley College

OtherTextClassification
Developed and trained a Convolutional Neural Network (CNN) for MRI brain tumor classification, annotating and labeling medical image data for supervised learning. Performed morphological preprocessing and labeling of brain regions, including ground-truth validation for training data. Published work and presented annotated results at research symposia and academic conferences. • Labeled and curated MRI slices for brain tumor detection • Validated and cross-checked data labels to ensure clinical accuracy • Used Python, TensorFlow/Keras for annotation and training cycles • Received research award for contributions to annotated medical data

Developed and trained a Convolutional Neural Network (CNN) for MRI brain tumor classification, annotating and labeling medical image data for supervised learning. Performed morphological preprocessing and labeling of brain regions, including ground-truth validation for training data. Published work and presented annotated results at research symposia and academic conferences. • Labeled and curated MRI slices for brain tumor detection • Validated and cross-checked data labels to ensure clinical accuracy • Used Python, TensorFlow/Keras for annotation and training cycles • Received research award for contributions to annotated medical data

2021 - 2022

Open Source Contributor — Persian Speech-to-Text

OtherAudioTranscription
Trained an automatic speech recognition (ASR) model for Farsi using large-scale annotated speech and text data from Common Voice. Performed data curation, preparation, and alignment of audio-text pairs to create a labeled dataset. Contributed a fully-trained model with aligned labeled data to the open-source Coqui STT repository. • Conducted audio transcription and text alignment for Persian ASR • Validated and preprocessed 271 hours of audio for model training • Used Python, Coqui STT, KenLM for speech-to-text labeling • Benchmarked and reviewed model outputs for quality and error metrics

Trained an automatic speech recognition (ASR) model for Farsi using large-scale annotated speech and text data from Common Voice. Performed data curation, preparation, and alignment of audio-text pairs to create a labeled dataset. Contributed a fully-trained model with aligned labeled data to the open-source Coqui STT repository. • Conducted audio transcription and text alignment for Persian ASR • Validated and preprocessed 271 hours of audio for model training • Used Python, Coqui STT, KenLM for speech-to-text labeling • Benchmarked and reviewed model outputs for quality and error metrics

Not specified

Education

U

University of California, Los Angeles

Bachelor of Science, Biology

Bachelor of Science
2022 - 2025
I

Irvine Valley College

Associate of Science, Biological Sciences

Associate of Science
2020 - 2022

Work History

R

Rossi

Computational Biologist

Los Angeles
2026 - Present
X

Xavvi

Head of AI Platform, Computational Biology & AI/ML

Los Angeles
2025 - Present