> Markdown version of [/jobs/ext/4583092-senior-data-scientist](https://www.wearedevelopers.com/jobs/ext/4583092-senior-data-scientist). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Senior Data Scientist - **Company:** The Bull - **Location:** United States (Remote available) - **Experience:** Expert - **Contract:** Permanent contract - **Skills:** Artificial Intelligence, Amazon Web Services, Computer Vision, Microsoft Azure, Code Review, Distributed Computing Environment, Python (Programming Language), Machine Learning, NoSQL, Software Product Management, SQL Databases, Web Analytics, Google Cloud, Graph Neural Networks, Pytorch, Retrieval-Augmented Generation, Large Language Models, Apache Spark, Llamaindex, Generative AI, Pandas, Pyspark, HuggingFace, Machine Learning Operations, GPT, Service Stack, Databricks - **Published:** November 5, 2026 - **Apply:** https://arc.dev/remote-jobs/j/redirect/pwttlnom0a ## About the Role Mandatory: * 5+ years building data science or ML solutions, with a track record of work that reached real users. * Python, SQL, NoSQL, and the standard analysis stack. * Breadth across ML domains - we'd like to see genuine hands-on depth in at least two of: LLMs and generative AI, computer vision, classical ML. * Production experience: you've deployed models - containers, pipelines, monitoring, retraining. You don't need to be an MLOps specialist, but you shouldn't need one to get something live. * Comfort in the cloud (Azure, AWS, or GCP) and with distributed data processing (Spark/Databricks or equivalent). * A strong mathematics or statistics background is preferred - you can reason why a model works, not just that it does. * Client-facing confidence: you can explain a technical trade-off to a non-technical stakeholder and push back when it matters. And you did this many times in the past. * Strong applied statistics: experiment design, uncertainty, calibration, and the discipline to distinguish a real effect from a hopeful one * Experience with modern AI tools, RAG architectures, vector databases, and embedding models * Familiarity with LLM frameworks and libraries (e.g., LangChain, LlamaIndex) * Experience evaluating ML systems against operational rather than academic criteria * Ability to communicate a negative result clearly to a non-technical audience and hold your position under pressure * You must speak Romanian natively and English advanced. Nice To Have * Exposure to regulated industries such as pharma, public sector, banking. * 2+ years of experience in managing teams - you have managed teams for building, deploying and maintaining an AI product/solution before. * Databricks experience * OCR experience * Graph Neural Networks * Fairness and bias auditing methodology * Familiarity with the EU AI Act's high-risk evaluation and documentation requirements * Domain exposure to public administration, law enforcement, migration or emergency services * Spatio-temporal modelling, survival analysis, or anomaly detection on log data ## Description As a Senior Data Scientist At Bull, You'll Lead The Technical Direction On Client Engagements And Turn Ambiguous Business Problems Into Working AI Products. You'll Set The Approach, Build The Core Of The Solution Yourself, And Bring The Rest Of The Team Along With You. Working In a Small Squad Of Engineers, Analysts, And PMs, You'll * Frame the problem: Sit with clients, cut through the ambiguity, and decide what's actually worth building up, including saying when ML isn't the answer. * Own the solution design: Choose the modelling approach and architecture and justify the trade-offs to both engineers and executives. * Raise the bar: Review code, mentor junior colleagues, and shape how we work across projects. * This job opens up for a specific long-term project that will require the following tasks: * Profile and assess data during feasibility - completeness, bias, label quality, whether the signal is there at all * Recommend continue or kill, with the evidence written down * Define the success metric at the gate, in terms the institution accepts and that will still mean something at scale * Build evaluation harnesses and baselines; benchmark models against them * Run bias and fairness testing, particularly where systems touch nationality, ethnicity or vulnerable groups * Measure pilot results across sites and quantify what the difference implies for wider rollout * Contribute to fundamental rights impact assessments alongside the governance function * Support the build itself, especially model iteration and error analysis Our Projects & Technology Stack * Real-world challenges across banking, insurance, manufacturing, pharmaceuticals, public sector and e-commerce, spanning call-centre logs, customer segmentation, IoT sensor streams, imagery, and web analytics. The scope is broad and keeps evolving as new data and technologies emerge. * Tech we love: Python, SQL, Pandas, PySpark, Databricks, MLflow, PyTorch, Hugging Face, and the LLM tooling ecosystem. We deploy on Azure, AWS, Google Cloud Platform and on-premises. * How we work: Small, focused teams (2-5 people), agile workflows, strong collaboration between engineers, analysts, scientists, and PMs.