Location: Remote within ±3 hours of the EAT time zone. 

Reports to: Associate Manager, Data 

About Pula 

At Pula, we are radically restructuring agricultural insurance, using technology to insure the previously unbanked, uninsured, untapped market of 1.5 billion smallholders worldwide. 

We work in over 14 countries across Africa and Asia, and by 2020, we have facilitated crop insurance cover to over 4.7 million farmers across Kenya, India, Rwanda, Uganda, Nigeria, Mali, Senegal, Zambia, Ethiopia, Madagascar, Tanzania, and Malawi. Since raising our Series A our operations have expanded globally including now Asia and Latin America. 

Pula is fast-paced, constantly adapting to new opportunities and challenges. We are a high performance, multi-cultural team. We expect our team to value performance, results and professionalism. 

At Pula, we are: 

•  Obsessed with results: We are responsible for our future and therefore we get things done!

•  Connect the Dots: We continuously identify different customer needs & business problems and build solutions that deliver value. 

•  Have a Can do Attitude! We dare to do things differently to make things work.

•  We have each others’ back: We look out for each other & we share high quality, timely feedback that makes us better. 

Position Overview 

As a Data Analyst (Data Engineering & AI Workflows) at Pula, you will clean, validate and move the field and geospatial data that powers our insurance products, farmer profiling and compliance projects. You will build and maintain the pipelines that take raw survey data from collection to delivery, and set up AI-assisted workflows that make data cleaning, quality checks and document extraction faster and repeatable. You will work closely with the Associate Manager, Data, the analyst team and the field teams across concurrent country projects. We are looking for someone who writes production-quality code, thinks in systems rather than one-off scripts, and documents their work so others can run it. 

Responsibilities 

Data cleaning and quality control 

•  Clean, standardise and validate large field datasets (farmer registrations, surveys, polygons, photos) collected through ODK and in-house tools. 

•  Run data quality checks on each batch following our gate-based DQC framework and SOPs, and resolve flagged issues with analysts and field teams. 

•  Reconcile collected data against master lists and reference data

•  Distinguish real data problems from defects in the checks themselves, and fix the checks when they are wrong. 

•  Identify gaps in current quality check SOPs and recommend improvements. 

Data engineering and pipelines 

•  Build and maintain ETL pipelines that move data from collection tools and PostgreSQL/PostGIS 

databases to analysis outputs and client deliverables. 

•  Schedule and monitor automated jobs (e.g. on cloud servers or GitHub Actions), with logging, incremental loads and clear failure alerts. 

•  Design and document data schemas, validation rules and data dictionaries for new country projects. 

•  Write clean, version-controlled, tested Python and SQL that other team members can run and maintain. 

•  Process geospatial data (field boundaries, GPS points, administrative layers) and support compliance checks such as EUDR polygon validation. 

AI workflows and automation 

•  Set up and maintain AI-assisted workflows using LLM tools (e.g. Claude, Claude Code, OpenAI or similar) for data cleaning, quality checks, report drafting and repetitive analyst tasks. 

•  Build document and image extraction pipelines (OCR and LLM-based) for sources such as land certificates, forms and field photos, with verification steps to measure accuracy.

•  Write and maintain prompts, reusable skills, project context files and runbooks so AI workflows give consistent, auditable results. 

•  Evaluate when AI is the right tool versus a deterministic script, and track cost, accuracy and data privacy risks of each approach. 

•  Train analysts to use AI tools safely and effectively in their daily work.

Analysis, reporting and improvement 

•  Use data analytics and project knowledge to identify risks and issues within projects, and propose practical solutions. 

•  Prepare dashboards, reports and delivery files for internal teams, partners and clients.

•  Develop new processes and procedures to improve project operations, and evaluate their effectiveness once implemented. 

•  Draft technical documentation, SOPs and training materials, and train analysts and field staff on new systems. 

•  Keep the manager informed through regular daily, weekly, and ad-hoc updates.

•  Demonstrate honesty and transparency, and communicate directly and openly, giving feedback constructively. 

Requirements 

Education and experience 

•  Bachelor’s degree in Computer Science, Statistics, Data Science, Information Systems, Economics, Geoinformatics, or a related field. 

•  Minimum of 2 years of experience in data analysis or data engineering, preferably in a fast-paced NGO or agri-tech environment. 

•  Demonstrated experience cleaning large, messy field or survey datasets end to end.

•  A portfolio, GitHub profile or work samples showing pipelines or automations you have built.

Technical skills 

•  Strong Python (required), including pandas and scripting for automation; R is a plus.

•  Strong SQL (required), with experience in PostgreSQL; PostGIS is a plus. 

•  Experience building and scheduling ETL pipelines, using Git, and working on Linux or cloud servers (e.g. AWS EC2). 

•  Hands-on experience using LLMs and AI tools in real workflows (e.g. Claude, ChatGPT, coding assistants, LLM APIs), including prompt design and checking AI outputs for accuracy. 

•  Familiarity with mobile data collection tools such as ODK, KoboToolbox, CommCare, or SurveyCTO. 

Nice to have 

•  Experience building agentic or multi-step AI workflows (e.g. Claude Code, MCP connectors, LangChain or similar), or running self-hosted/open-source models. 

•  Experience with OCR and document extraction (e.g. Tesseract, PaddleOCR, cloud vision APIs).

•  Experience with geospatial tools (GeoPandas, QGIS, Google Earth Engine) or EUDR/deforestation compliance data. 

•  Experience with orchestration tools (Airflow, Prefect, GitHub Actions) or Docker.