Satori Analytics

**What Your Day Might Look Like:** - **Segment the customers:** Design segmentation solutions from customer, transactional, and behavioural data using clustering techniques such as K-Means, Gaussian Mixture Models, or DBSCAN. - **Engineer the signal:** Build customer-level features — recency, frequency, monetary value, spend trends, engagement, and lifecycle indicators — and evaluate results through metrics, stability analysis, and business interpretability. - **Explain the "why":** Use explainability techniques, particularly SHAP, to understand model behaviour and translate feature importance into clear customer narratives, profiles, and personas. - **Go beyond segments:** Contribute to targeting, campaign analytics, propensity modelling, churn prediction, and other marketing use cases, surfacing patterns and performance drivers in customer and marketing KPIs. - **Align with the business:** Work with Marketing and BI to define meaningful KPIs and connect analytical outputs to business objectives. - **Tell the story:** Present findings to technical and non-technical stakeholders through clear visualisations and business language. - **Work clean:** Write reusable, reproducible, well-documented code, collaborating with Data Engineers and BI to prepare and validate datasets.

**What Your Day Might Look Like:** - **Build the models:** Develop and enhance Marketing Mix Models to estimate the impact of media, promotions, pricing, seasonality, and other business drivers on performance. - **Quantify what matters:** Apply regression, time-series, econometric, and ML techniques to measure incremental impact — modelling carryover, saturation, diminishing returns, and response curves. - **Optimise the spend:** Develop scenario-planning and optimization approaches to guide media budget allocation and investment decisions. - **Interrogate the results:** Evaluate assumptions, uncertainty, and business plausibility rather than relying on statistical fit alone, using SHAP, diagnostics, and sensitivity analysis to explain the drivers. - **Tell the story:** Translate outputs into clear recommendations on channel performance, ROI, and budget strategy for both technical and non-technical audiences. - **Partner across the business:** Work with Marketing, Commercial, Finance, BI, and Data Engineering to define questions, KPIs, and success criteria, and to operationalise clean, reproducible workflows. - **Raise the bar:** Support less-experienced colleagues and contribute to reusable methodologies and Data Science best practices.

• Connect the framework: Integrate a modular simulation optimisation framework with production AnyLogic models, databases, APIs, and plant-data environments. • Turn questions into use cases: Translate operational questions — scheduling, resource allocation, operating parameters, energy, logistics, throughput — into simulation and optimisation use cases. • Build in Java: Develop and maintain Java-based components that execute scenarios, marshal inputs, parse outputs, and enable reliable optimisation runs. • Make results credible: Define and validate simulation input/output data contracts and business KPIs, ensuring recommendations are reproducible and usable by operational stakeholders. • Optimise smartly: Apply the right optimisation approaches for real-valued, scheduling, selection, and multi-objective problems — and design efficient experiments for long-running simulations. • Prove the value: Validate recommendations against live models, quantify operational benefit, and support the transition from prototype to a repeatable operational process. • Operationalise it: Contribute to dashboards, recurring study scheduling, workflow integration, technical documentation, and knowledge transfer. • Collaborate closely: Work with client stakeholders and the wider delivery team to prioritise delivery, resolve issues, and maintain technical quality.

**What Your Day Might Look Like:** - Design, build, and maintain data pipelines that serve models, data, and AI workflows to internal and client-facing applications. - Work across database types — relational, vector, and graph — modelling and storing data appropriately for each access pattern, in close collaboration with data scientists. - Build and maintain dbt models — writing transformation logic, tests, and documentation that ensure data quality and traceability end-to-end. - Operate what you build: instrument pipelines with logging, metrics, and tracing; diagnose and resolve production data issues before they become someone else's problem. - Write clean, tested, production-quality code and contribute to CI/CD pipelines and infrastructure-as-code. - Show up to code reviews, design discussions, and retrospectives — and have something worth saying.

• Architect and build production data pipelines and data platforms that serve models, data, and AI workflows to internal and client-facing applications — and stay accountable for them under live conditions. • Own non-functional quality across your domain: latency and throughput budgets, scalability, reliability, observability, and cost. • Lead the design and operation of multi-model data stores — relational (PostgreSQL, MySQL), vector (Pinecone, Weaviate, pgvector), and graph (Neo4j, Neptune) — applying the right tool to each access pattern, not the most fashionable one. • Set technical direction: write design docs, make build-vs-buy calls, and defend your approach with evidence rather than instinct. • Work across the stack when the problem demands it — services, data access, infrastructure-as-code, CI/CD — and diagnose it when things drift in production. • Raise the floor for the team: mentor mid-level and junior engineers, run rigorous code reviews, and hold the quality bar without making it someone else's job to ask you.

**What Your Day Might Look Like:** - **Cloud infrastructure as code**: Own and extend our Terraform estate across multiple GCP environments (base, core, obs, dev, test, prod), including GKE clusters, Cloud SQL (Postgres/MySQL), networking, buckets, and IAM. Drive the in-progress "Neo" platform rollout and the cutover/retirement of legacy infrastructure. - **Kubernetes & containers**: Manage workloads on GKE, maintain Dockerfiles and Helm-style application configs for ~10 backend services, and tune autoscaling, resource limits, and pod disruption budgets. - **Maintain and improve our GitHub Actions pipelines**: PR checks (Python/JS lint, type-check, tests), Terraform prechecks, image builds and pushes, auto-deploy, and DB-migration labelling/gating. Reduce build times and flakiness, and make deploys self-service for product teams. - **Data & messaging infrastructure**: Operate Postgres, Redis, and Celery-based async workers; manage Alembic migrations, queue health, and backpressure for long-running simulation jobs. - **Observability**: Own our monitoring stack — Grafana dashboards, ClickHouse, Langfuse (LLM tracing), and Celery queue metrics. Build alerting and SLOs so we catch issues before customers do. - **Security & secrets**: Manage secret distribution, least-privilege IAM, and remediation tracking. Partner with engineering on findings in our security assessment process. - **Cost & reliability**: Keep an eye on cloud and LLM-proxy (LiteLLM) spend, right-size resources, and improve resilience of the simulation and evaluation pipelines. **You'll work with:** - Cloud: Google Cloud Platform (GKE, Cloud SQL, GCS, IAM); some AWS / IBM footprint - IaC: Terraform (>= 1.14), multi-environment root modules - Containers/orchestration: Docker, docker compose (local), Kubernetes / GKE - CI/CD: GitHub Actions - Backend: Python 3.13+ (managed with uv), Celery, FastAPI-style HTTP APIs; Node/Express services - Data: PostgreSQL, MySQL, Redis, ClickHouse - Observability: Grafana, Langfuse, custom Celery metrics - LLM infra: LiteLLM proxy

• Build and maintain the services, data models, and APIs that power the platform — designed for correctness, testability, and scale. • Work on the systems that coordinate complex, multi-step interactions between AI agents and external systems, improving their reliability and throughput. • Design systems that grade agent outputs, combining deterministic checks with model-assisted judgment — and make scoring reliable, explainable, and reproducible. • Build pipelines that generate, transform, and quality-check large volumes of structured data and benchmark content. • Add the tests, instrumentation, and safeguards needed to trust outputs from systems that are inherently non-deterministic.

Role Description Are you passionate about AI? 🤖 At Satori Analytics, we aim to change the world one algorithm at a time by bringing clarity to global brands through Data & AI. From cloud-based ecosystems for fintech to predictive models for airlines, our cutting-edge solutions cover the entire data lifecycle—from ingestion to AI applications. As a fast-growing scale-up, our team of 100+ tech specialists—including Data Engineers, Data Scientists, and more—delivers innovative analytics solutions across industries like FMCG, retail, manufacturing and FSI. Join us as we lead the data revolution in South-Eastern Europe and beyond! Together with a partnering company, we're looking for a DevOps / Platform Engineer to own and evolve the infrastructure that keeps this platform reliable (AI agent evaluation platform), observable, secure, and fast to ship to. You'll work closely with backend, ML, and frontend engineers to make deploying and operating services boring, repeatable, and safe. What Your Day Might Look Like: - Cloud infrastructure as code: Own and extend our Terraform estate across multiple GCP environments (base, core, obs, dev, test, prod), including GKE clusters, Cloud SQL (Postgres/MySQL), networking, buckets, and IAM. Drive the in-progress "Neo" platform rollout and the cutover/retirement of legacy infrastructure. - Kubernetes & containers: Manage workloads on GKE, maintain Dockerfiles and Helm-style application configs for ~10 backend services, and tune autoscaling, resource limits, and pod disruption budgets. - Maintain and improve our GitHub Actions pipelines: PR checks (Python/JS lint, type-check, tests), Terraform prechecks, image builds and pushes, auto-deploy, and DB-migration labelling/gating. Reduce build times and flakiness, and make deploys self-service for product teams. - Data & messaging infrastructure: Operate Postgres, Redis, and Celery-based async workers; manage Alembic migrations, queue health, and backpressure for long-running simulation jobs. - Observability: Own our monitoring stack — Grafana dashboards, ClickHouse, Langfuse (LLM tracing), and Celery queue metrics. Build alerting and SLOs so we catch issues before customers do. - Security & secrets: Manage secret distribution, least-privilege IAM, and remediation tracking. Partner with engineering on findings in our security assessment process. - Cost & reliability: Keep an eye on cloud and LLM-proxy (LiteLLM) spend, right-size resources, and improve resilience of the simulation and evaluation pipelines. Qualifications - 3+ years in DevOps / SRE / Platform Engineering, or strong backend experience with heavy infra ownership. - Solid hands-on Terraform (modules, state, multi-environment) and cloud experience (GCP preferred; AWS/Azure transferable). - Production Kubernetes experience: deployments, services, autoscaling, debugging pods, rollouts/rollbacks. - Strong Docker fundamentals and comfort writing/optimising Dockerfiles. - CI/CD pipeline design and maintenance (GitHub Actions, or equivalent like GitLab CI / CircleCI). - Comfortable scripting and reading code in Python and/or Bash; able to navigate a polyglot monorepo. - Operational experience with relational databases and managed database services (migrations, backups, performance). - A reliability mindset: monitoring, alerting, incident response, and writing runbooks. Requirements - Experience operating Celery / distributed task queues and Redis at scale. - Familiarity with LLM/AI infrastructure (model proxies, GPU scheduling, token/cost management). - Observability tooling depth (Grafana, Prometheus, ClickHouse, OpenTelemetry, Langfuse or similar tracing). - Security/compliance experience (IAM hardening, secret management, vulnerability remediation). - Cost-optimisation experience for cloud + third-party API spend. - Experience supporting a monorepo with multiple language ecosystems and editable/internal package dependencies. Benefits - Competitive salary. - Training budget to level up your skills from top tech partners like Microsoft, AWS, Salesforce, and Databricks – whether it’s certifications or courses, we’ve got you covered. - Private insurance, top-tier tech gear, and the chance to work with a stellar crew. Company Description Ready to create some data magic with us? Hit that apply button and let’s get started. ✨

Role Description As a fast-growing scale-up, our team of 100+ tech specialists—including Data Engineers, Data Scientists, and more—delivers innovative analytics solutions across industries like FMCG, retail, manufacturing and FSI. Together with a partnering company, we're looking for a strong Software Engineer to build the systems behind an AI agent evaluation platform — systems that test, grade, and stress AI agents at scale, and turn the results into actionable signal about how those agents behave. This is a software engineering role first — designing services, data models, and pipelines that have to be correct, tested, and maintainable — applied to an AI problem domain. You'll work with LLMs regularly, but the core of the job is engineering, not prompt-tuning. What Your Day Might Look Like: - Backend services & APIs: Build and maintain the services, data models, and APIs that power the platform — designed for correctness, testability, and scale. - Simulation & orchestration: Work on the systems that coordinate complex, multi-step interactions between AI agents and external systems, improving their reliability and throughput. - Evaluation & scoring: Design systems that grade agent outputs, combining deterministic checks with model-assisted judgment — and make scoring reliable, explainable, and reproducible. - Data pipelines: Build pipelines that generate, transform, and quality-check large volumes of structured data and benchmark content. - Quality & reliability: Add the tests, instrumentation, and safeguards needed to trust outputs from systems that are inherently non-deterministic. Qualifications - 4+ years building and shipping production software, with strong proficiency in Python. - Deep software engineering fundamentals: system and API design, data modeling, concurrency/async, testing strategy, debugging, and code review. - Experience designing and operating distributed or service-oriented systems (queues, workers, APIs) — not just calling them. - Comfort designing schemas and working with relational databases, plus the migrations and performance concerns that come with them. - Working knowledge of LLM APIs — orchestration, structured outputs, and handling non-determinism. - Ability to reason about correctness of probabilistic systems: how to test, measure, and trust outputs that aren't byte-for-byte deterministic. - High quality bar: you write tests, types, and docs by default, and you keep changes small and reviewable. Requirements - Bonus points for experience building agentic or multi-agent systems, tool-use, or orchestration frameworks. - Background in evaluation / benchmarking of ML or LLM systems (rubrics, golden datasets, model-as-judge, inter-rater reliability). - Experience with distributed task queues and async workloads. - Modern Python tooling and typed codebases (e.g. type checkers, linters, Pydantic, FastAPI). - Retrieval / search experience and working with data ingest pipelines. - Some comfort with the infra side (Docker, CI/CD) so you can ship what you build. Benefits - Competitive salary. - Training budget to level up your skills from top tech partners like Microsoft, AWS, Salesforce, and Databricks – whether it’s certifications or courses, we’ve got you covered. - Private insurance, top-tier tech gear, and the chance to work with a stellar crew.

**What Your Day Might Look Like:** - **Ship product features:** Build chat interfaces, admin dashboards, authentication flows, and document workflows that enterprise users rely on every day. - **Own the full lifecycle:** Contribute across the full frontend stack, from discussing requirements with product, to writing typed components, to shipping and monitoring in production. - **Translate AI into UI:** Turn LLM capabilities (streaming responses, citations, voice input, multi-agent flows) into intuitive, trustworthy interfaces. - **Set the quality bar:** Shape frontend conventions the team follows: accessibility, performance, testing patterns, and component design. - **Keep systems healthy:** Monitor and troubleshoot production with Azure Application Insights and structured Winston logs to ensure reliability and great UX. - **Collaborate across the stack:** Work closely with Python backend engineers, product stakeholders, and designers to turn use cases into working features. - **Document and share:** Maintain clear component documentation and architectural notes to keep the team aligned as the products grow. - **Agile team structure:** Participate in agile methodology ceremonies and catch up with your team regularly.