описание
About Conductor Conductor is the leading enterprise AEO platform. Today’s top enterprise brands use Conductor to grow authority and visibility in both AI and traditional search engines. From tracking visibility in LLMs to real-time monitoring of technical site health and scaled AI content creation, Conductor provides a single source of truth that fuels digital growth—all from one platform. We’re looking for a Staff-level engineer who can take ambiguous problems from exploration to production.
задачи
- Design and ship backend services, APIs, and data pipelines that power AI products, integrating ML models, LLMs, and internal and external systems.
- Make pragmatic architectural decisions across application, data, model, and infrastructure boundaries, with reliability, observability, performance, cost, and simplicity in mind.
- Own substantial workstreams end to end: exploration, architecture, implementation, deployment, and operation.
- Develop and evaluate ML algorithms, statistical methods, ranking and clustering systems, and retrieval pipelines for problems where the solution isn’t known in advance.
- Analyze noisy internal datasets to diagnose product behavior and find opportunities.
- Design experiments and metrics that compare approaches rigorously, prototype quickly, and carry what works into maintainable production systems.
- Read research critically and adapt it to real product constraints.
- Create reusable, reproducible evaluation pipelines for models, prompts, retrieval, and agents, rather than relying on manual inspection.
- Set up experiment tracking, dataset and artifact versioning, and monitoring, so engineers can compare approaches and investigate failures.
- Help decide when conventional software, statistics, classical ML, or LLMs is the right tool.
- Containerize, deploy, and operate your services, and build the CI/CD and operational tooling around them.
- Diagnose failures across application, data, model, and infrastructure layers.
- Reason about latency, throughput, storage, reliability, and cost.
- You measure rather than argue from intuition, and you challenge assumptions, including your own, when the evidence says otherwise.
- You can go deep without turning every problem into a research project, and you know when to prototype and when to engineer properly.
- You treat observability, testing, and evaluation as part of the system.
- You document your reasoning so others can understand it later, and you communicate clearly with both specialists and non-specialists.
- After ramping up, you own major technical workstreams with minimal supervision.
- You work with other senior engineers as a high-bandwidth peer: splitting hard problems, pursuing separate lines of investigation, challenging each other’s designs, and combining the results into systems that work.
требования
- Staff-level experience, or an exceptionally strong Senior engineer approaching it.
- A track record of shipping non-trivial, high-impact projects more than once, ideally including at smaller companies.
- The ability to structure ambiguous problems, separate essential from accidental complexity, and work effectively with little supervision, coaching, or task breakdown.
- Deep production Python experience, not only notebooks or prototypes.
- Strong command of tools such as NumPy, pandas, SciPy, scikit-learn, PyTorch, Pydantic, and FastAPI, and comfort profiling, testing, and optimizing Python systems.
- Hands-on experience building, evaluating, or operating real ML or statistical systems, beyond calling pretrained model APIs.
- Solid grounding in probability, statistics, optimization, and experimental design, plus the ability to judge whether an unfamiliar algorithm or paper fits a problem.
- Experience separating real signal from confounders, measurement artifacts, and evaluation failure modes.
- A research degree is welcome but not required. Demonstrated, serious work in ML and experimentation matters more.
- Experience designing backend or distributed systems and data pipelines at meaningful scale.
- Strong grounding in APIs, async processing, queues, databases, failure handling, and observability.
- Practical experience with Docker, Kubernetes, AWS (e.g., S3, SQS, Lambda, ECS/EKS, DynamoDB, CloudWatch), CI/CD, and a workflow orchestrator such as Airflow, Dagster, or Prefect.
- Будет плюсом: Experience in a research-intensive or deep-tech setting, such as robotics, computer vision, search and ranking, recommenders, scientific computing, optimization, autonomous systems, or computational biology; Building or materially modifying ML algorithms, or designing evaluations where no off-the-shelf benchmark exists; Semantic search, embeddings, or vector databases, and LLM applications such as agents, RAG, or structured generation; Experiment-tracking and evaluation tools such as Weights & Biases or MLflow, or custom internal tooling; Event-driven architectures and Kubernetes-based ML workloads; Moving experimental ML systems into reliable production services
условия
- Conductor maintains competitive, performance-based compensation programs.
- The Berlin base salary range for this role is currently 120,000 - 130,000 euro.