Joaquin Salvador Machulsky

AI & ML Engineer | Data Scientist | LLM Specialist

Building production AI systems with PyTorch, LLMs, and cloud infrastructure.

JM
Joaquin Machulsky

About Me

AI & ML Engineer specializing in production-grade AI systems and LLM solutions.

My Background

Data Science professional with focus on machine learning and AI systems.

My Approach

Building scalable AI solutions that transform data into actionable insights.

Continuous Learning

Staying current with the latest in ML, LLMs, and AI development.

Technical Skills

A comprehensive toolkit of technologies and frameworks for tackling any data science challenge

Programming Languages

PythonSQLC++JavaScriptHTML

AI/ML & LLMs

Gemini LLMRAG SystemsSemantic SearchEmbeddingsPyTorchTensorFlowScikit-learnComputer VisionNLP

Data & Backend

PandasNumPyOpenCVFlaskSQLAlchemyMarshmallowSeleniumSciPyMatplotlibSeaborn

Cloud & DevOps

AWS (EC2, S3, IAM, RDS)Elastic BeanstalkCodePipelineDockerGoogle Cloud APIsPostman

Databases

PostgreSQLMySQLSQL ServerVector SearchJSON-based storage

Tools & Collaboration

GitGitHubReactMaterial UIJupyterVS Code

Specializations

LLM Operations

Production-grade LLM systems and RAG architectures

AI Agent Development

Building scalable AI agents and debate mechanisms

Computer Vision

PyTorch-based models for medical imaging and detection

Time Series Analysis

Forecasting models for market analysis and trends

Featured Projects

A selection of projects demonstrating my expertise in data science, machine learning, and end-to-end AI solution development

AI Agents Alignment: Scalable AI Safety via Debate

AI Agents Alignment: Scalable AI Safety via Debate

Research exploring debate mechanisms for AI alignment. Implemented full simulation game (zero-sum) of a debate protocol using PyTorch, including computer vision models, turn-based strategies, and agent asymmetry capabilities.

PythonPyTorchComputer VisionGame TheoryAI Safety
Semantic Search RAG System

Semantic Search RAG System

Production-grade semantic search Agentic RAG system using Gemini LLM and Google embeddings, enabling natural language queries for a 544-product catalog. Deployed on AWS with high availability.

PythonGemini LLMRAGGoogle EmbeddingsAWSVector Search
Melanoma Detection Computer Vision Model

Melanoma Detection Computer Vision Model

PyTorch-based computer vision model for melanoma detection achieving 80% recall. Includes comprehensive model evaluation, validation techniques, and performance optimization.

PythonPyTorchComputer VisionOpenCVMedical Imaging
Dream Journal NLP Analysis

Dream Journal NLP Analysis

Web scraping and NLP-based analysis of dream narratives, including sentiment analysis, coherence evaluation, and topic modeling. Investigated relationships between dreams and lottery numbers using statistical analysis.

PythonNLPWeb ScrapingSentiment AnalysisTopic Modeling

Want to see more projects?

View all on GitHub

Experience & Education

My academic and professional journey in the world of data science

Professional Experience

Data Scientist

NETV S.A.

August 2024 - Present
Remote
PythonPyTorchGemini LLMRAGAWSDockerFlaskPostgreSQL

Freelance ML & Data Developer

LAS MERCEDES

March 2025 - Present
Remote
PythonFlaskAWSPostgreSQLPandasScikit-learn

Data Analyst & Data Developer

RUSSO SEGUROS

November 2023 - May 2025
Buenos Aires, Argentina
PythonPostgreSQLPandasSeleniumMatplotlibSeaborn

Education & Certifications

Master in Data Science (Artificial Intelligence)

University of Buenos Aires

2023 - 2025
Buenos Aires, Argentina

Bachelor in Data Science

University of Buenos Aires

2020 - 2023
Buenos Aires, Argentina

English Proficiency (FCE B2)

Cambridge English

Certified
International

Let's Connect!

I'm always open to new opportunities, collaborations, or simply a chat about data science. Don't hesitate to reach out!

Looking for a Data Scientist?

I'm available for full-time opportunities, freelance projects, research collaborations, or mentoring. My expertise includes LLM systems, RAG architectures, computer vision, and end-to-end ML development.

Built with ❤️ using Next.js, TypeScript and Tailwind CSS

© 2025 Joaquin Salvador Machulsky. All rights reserved.