Machine Learning Engineer

MercadoLibre
Buenos Aires
Workplace: HybridFull timeFunction: Data Science & Machine LearningSkills: []

Design and scale low-latency inference solutions for Computer Vision and NLP, taking LLM/VLM systems from exploration and context engineering to agentic, multimedia reasoning workflows. Build robust evaluation pipelines to measure impact and prevent regressions, and continuously improve models using prompt learning and data-driven iteration. Define component contracts using spec-driven development, and develop the data pipelines for training, fine-tuning, and ongoing evaluation.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
MercadoLibre
MercadoLibre
1 week ago

Machine Learning Engineer

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 5 hours agoStatus: Live

Job Summary

Design and scale low-latency inference solutions for Computer Vision and NLP, taking LLM/VLM systems from exploration and context engineering to agentic, multimedia reasoning workflows. Build robust evaluation pipelines to measure impact and prevent regressions, and continuously improve models using prompt learning and data-driven iteration. Define component contracts using spec-driven development, and develop the data pipelines for training, fine-tuning, and ongoing evaluation.
Location: Buenos Aires
Workplace: Hybrid
Employment Type: Full time
Job Function: Data Science & Machine Learning

Key Responsibilities

  • •Design and evolve low-latency inference solutions for Computer Vision and NLP, optimizing throughput and resource usage for live experience.
  • •Own the end-to-end lifecycle of LLMs/VLMs, from exploration and context engineering to implementation of agentic workflows that reason over multimedia content.
  • •Ensure model quality by building reproducible eval pipelines to measure the impact of model/prompt/flow changes and mitigate regressions and hallucinations.
  • •Iterate systematically using prompt learning and continuous data-driven iteration to maximize recall and precision.
  • •Define system components and contracts with spec-driven development, and build data pipelines needed for training, fine-tuning, and ongoing evaluation.

Key Requirements

  • •Master end-to-end lifecycle of Computer Vision and NLP models, including training, evaluation, and inference.
  • •Optimize and serve low-latency models using PyTorch and Hugging Face, with serialization formats like ONNX and high-performance inference engines such as vLLM or Llama.cpp.
  • •Have real experience with LLM-based solutions including Context Engineering, RAG, and systematic Prompt Iteration / Prompt Learning.
  • •Develop agentic flows and complex orchestration using LangChain, LangGraph, or the Google SDK, plus evaluation metrics with LLM eval frameworks (e.g., LangSmith, Ragas).
  • •Use efficient fine-tuning methods like QLoRA and build data/engineering infrastructure with Python (clean architecture, testing, CI/CD) and massive pipelines using BigQuery and advanced SQL.
Experience:Computer VisionNLPLLMsRAG
Languages:Spanish
Tech Stack:PyTorchHuggingFaceONNXVLLMLlama.cppLLMsVLMsContext EngineeringRAGPrompt IterationPrompt LearningLangChainLangGraphGoogle SDKLangSmithRagasEvalsQLoRAPythonClean Architecture

Company Brief

MercadoLibre
Operates Latin America's leading e‑commerce marketplace and fintech ecosystem, offering online buying and selling, classifieds, payment processing, credit, and logistics services across multiple countries in the region.
Industry: Online Marketplaces
Company Size: Enterprise (1,001+ employees)
Revenue: USD 1B+
Growth: Public Company
Valuation: Public Company (Market Cap in USD)
Funding: IPO / Publicly Listed
Headquarters: Buenos Aires, Argentina
Founded: 1999
Glassdoor
Glassdoor: 3.8
WebsiteLinkedIn