Data Scientist

The company builds and operates production AI systems for companies ranging from seed-stage startups to Meta. We are hiring a Data Scientist to own the part that comes before the system exists: what question is actually being asked, which method answers it, whether the answer holds up, and how anyone would know.The role is fully remote across Latin America, aligned to your client's working day.

You will not be handed a specification. The company has shipped models into production since 2016, and the work here starts with a client who knows what bothers them and does not yet know what to measure. Turning that into a model, a baseline, and a number someone can act on is the job.

Where this role sits

The company's engineering organization is built around four lanes. The Data Engineer lane owns pipelines, storage, and the retrieval layer. The AI Engineer lane owns production behavior. The Software Engineer lane owns AI-augmented product delivery. This role is the Data Scientist lane, and it owns the question and the method.

One question places the boundary: when the output is wrong, whose problem is it? "The method was inappropriate, or the question was the wrong one" is yours . "The system did the wrong thing with an appropriate method" is the AI Engineer's.

Not quite your profile? Check our other openings

• If you build the pipelines and retrieval layer models depend on — Data Engineer
• If you own how an AI system behaves in production — AI Engineer
• If you ship product software with agents in your toolchain — AI-Augmented Software Engineer
• If you've done all of the above and answered to the client directly — Forward Deployed Engineer

What you will build

• The question. Translating a business decision into something measurable: what would count as success, what the baseline already achieves, and what a useful improvement over it would look like.
• Method and model. Supervised and unsupervised methods, deep learning with CNNs and Transformers in PyTorch, and the judgment to choose between them, including the cases where a simpler model or no model at all is the right answer.
• Fine-tuning and adaptation. Adapting pretrained and open-weight models with LoRA, QLoRA or PEFT, and a clear view of when fine-tuning is the wrong answer.
• Measurement and experimental design. Baselines, holdouts, offline and online evaluation, and honest reporting of uncertainty. If you say the model is better, you should be able to say better than what, by how much, and how sure you are.
• Error analysis and handover. Telling apart a data problem, a label problem, a method problem and a framing problem, and handing the model over packaged, documented, and with the metric that defines whether it is still working.
• Work inside the client's environment. Their repositories, their standups, sometimes their customer calls. The company is SOC 2 certified, client code stays in client repositories, and some engagements carry additional requirements such as HIPAA.

How we work

Our engineers build with AI every day. Claude Code, Codex, and similar tools are part of the standard toolchain here, not an experiment. We run an automated audit across the whole codebase on day one and every day after, grading security, cost, and architecture findings by severity with the exact file and line, so a small team can move quickly without quality drifting. We stay vendor-neutral across OpenAI, Anthropic, and open-weight models, and we run Valkyrie, our own production layer, when a single interface to any model is the right call.

About the company

The company is a San Francisco based software development company that has been building intelligent applications since 2016. We provide nearshore AI engineering teams to organizations that need production AI faster than they can hire for it: as an embedded engineering team, as AI staff augmentation alongside an existing team, or as a full project build. Our engineers work from Latin America, aligned to United States time zones, and have delivered for Twitter, Meta, Discovery Channel, Omnicom, UnitedHealth, and CENTEGIX.

We hire for seniority and test for it before anyone joins a client team. We support engineers in going deep on the modern AI stack, and we give time back to open-source work, community teaching, and philanthropy.

Apply at [link for members] or write to us at [link for members]

Requirements

Basic qualifications

• 5+ years building and validating models against real decisions, with Python as your primary language, plus the engineering fundamentals that go with it: testing, code review, Git, containers, and version control for data and experiments.
• Depth in machine learning and deep learning: supervised and unsupervised methods, CNNs and Transformers, with hands-on PyTorch or an equivalent framework, including training, custom modules, and optimization.
• Expert command of pandas, NumPy and scikit-learn.
• Strong foundation in statistics and experimental design: baselines, holdouts, significance, and the ability to state how confident you are and why.
• Fine-tuning or adaptation of pretrained models, and a clear view of when fine-tuning is the wrong answer.
• Error analysis as a discipline. You can separate a data problem from a label problem, a method problem and a framing problem, rather than reaching for a larger model.
• Cloud experience with AWS, GCP or Azure and their AI services such as Amazon SageMaker, Google Vertex AI or Azure Machine Learning.
• Active use of AI-assisted coding tools such as Claude Code, Cursor, or GitHub Copilot in real delivery work.
• Clear written and spoken English, C1 or above, and the confidence to explain a technical trade-off directly to a client.
• Bachelor's or Master's degree in Computer Science, Data Science, or a related field, or equivalent professional experience.

Preferred qualifications

• Evaluation of LLM systems: test-set design, model-as-judge or equivalent scoring, and benchmark selection you can defend.
• Open-weight fine-tuning and adaptation at depth: LoRA, QLoRA, PEFT, and the cost and latency trade-offs involved.
• Multimodal work covering vision, speech, or document understanding alongside text.
• Causal inference, uplift modeling, or forecasting at scale.
• Experiment tracking and MLOps tooling: MLflow, Weights & Biases, Kubeflow or similar.
• Delivery under a compliance regime such as SOC 2 or HIPAA.
• Contributions to open-source AI libraries, published research or technical writing, or active participation in the AI community.

Benefits

• Paid time off (PTO)

• U.S. Holidays

• AI Training

• Mentored career development

• Profit sharing

• $US remuneration