Why Reelo ?
In today’s customer centric world, it’s important for every business to market like the best brands in the world. As competition for customers’ attention increases, it has never been more important for small and mid-market F&B and retail businesses to have 360° view of the people that keep them in business. But collecting, understanding and taking action on customer data is riddled with market, resource, and technology challenges. Solving these challenges is why we get up in the morning.
Reelo is the new-age customer marketing platform for restaurant & retail businesses to grow their revenue, get more customers and build a stronger brand, in an incredibly simple manner.
We understand that it's easier than ever to start a restaurant & retail business, it’s harder than ever to grow it. Specifically, most of these businesses have two challenges with growth:
1. Getting people to make their first purchase.
2. Getting those people to come back and buy again (and again and again).
These businesses and their teams already have 1,000 things to do — wrestling with overly complicated marketing tools doesn’t need to get added to the list.
Our SaaS platform makes it fast & easy for businesses to understand more about their customers and continously engage with them in a smart & automated way - making them look like a professional without the effort.
There is a massive opportunity in front of us. We’re building a business that is big, meaningful, and lasting. For that, we’re building a strong, diverse team of curious, creative people who want to find purpose in their work and support each other in the process. If that sounds like you, we’d love to hear from you.
About the role
We're hiring an SDE (Insights) to build the analytics features our customers use to run their business. You'll write the queries behind them and make those queries fast. Reelo is a customer retention platform for restaurant and retail brands, covering loyalty, feedback, customer segments and WhatsApp/SMS/email campaigns.
Our customers ask questions like: who are my best customers, who stopped visiting, and did that campaign bring people back? You'll turn those questions into dashboards, reports, segments and metrics, built on ClickHouse and MongoDB data with millions of orders, customers and messages.
This is a development role focused on data. You won't manage servers, deployments or cloud cost. The platform team owns those. Your job is correct numbers, fast queries and insights people act on.
What you'll do
- Build insight features end to end. Customer and store dashboards, campaign performance reports, loyalty reports (points earned, redeemed and expired), feedback analytics and exports. You'll write the query and the Node.js API that serves it.
- Write analytical queries. ClickHouse SQL for aggregates, time series, cohorts, funnels and window functions. MongoDB aggregation pipelines where data still lives there.
- Build customer segments. RFM, recency and frequency buckets, lapsed and at-risk customers, top spenders, and audience recommendations for campaigns. Each must be exact, so the right customers get the message.
- Make slow queries fast. Read query plans and system.query_log. Fix sort keys, filters, joins and pre-aggregations, and prove it with before/after timings and rows read.
- Model data for analytics. Design ClickHouse tables, materialized views and roll-ups (AggregatingMergeTree, argMax, uniqExact) that answer dashboard questions without scanning raw data every time.
- Keep the numbers right. Reconcile metrics between MongoDB and ClickHouse. Handle timezones, date boundaries, duplicates and late-arriving data. Write tests so a metric can't silently change.
- Work with product and customer success. Turn a vague question ("why did repeat visits drop?") into a precise metric, a query and a clear answer.
What you'll work with
You'll spend most of your time in SQL and aggregation pipelines. The infrastructure under them is run by the platform team.
| Area | What we use | How much you'll use it |
| Analytics database | ClickHouse Cloud: MergeTree family, materialized views, dictionaries, @clickhouse/client | Every day |
| Transactional database | MongoDB with Mongoose: aggregation pipelines, indexes | Often |
| API layer | Node.js and Express.js: endpoints that serve your queries to dashboards | Often |
| Local analysis | chDB, Jupyter, Parquet and CSV exports | Sometimes |
| Quality | Jest tests for queries and metrics, ESLint, Prettier | Every change |
| Out of scope | Deployments, AWS infrastructure, servers, cloud cost | Owned by the platform team |
What you bring
- 2+ years building data-heavy software, as a backend, analytics or data engineer, with SQL as your main tool.
- Strong SQL. You write window functions, CTEs, conditional aggregates and multi-level GROUP BY without looking them up. You know where JOINs fan out and double-count.
- Query optimisation. You've made a slow query fast and can explain why it worked: sort or index keys, partition pruning, reading fewer columns, pre-aggregation, or a better join order.
- A columnar database in production. ClickHouse is ideal. BigQuery, Redshift, Snowflake, Druid or DuckDB also count if you understand why columnar queries behave differently.
- MongoDB aggregation pipelines. $match before $group, $lookup costs, and using indexes inside a pipeline.
- Analytics thinking. Cohorts, retention curves, funnels, RFM, customer lifetime value, churn, repeat rate, averages vs medians, and percentiles.
- Statistics basics. Distributions, outliers, z-scores, sampling, and knowing when a change is noise rather than a real trend.
- Care about correct numbers. You reconcile a metric against a second source before you ship it, and you handle timezones, nulls and duplicates on purpose.
- Working JavaScript (Node.js). Enough to write and test the API endpoint around your query; deep backend experience isn't required.
- Clear explanations. You can show a non-technical stakeholder what a number means and what it doesn't.
AI-assisted development
You'll use Claude Code every day to write queries, explore data and build features faster. We want someone who already uses it well on real data work.
How we expect you to use them:
- Check every number. Claude can write a query that runs and is still wrong. You check results with a second query, a known total or a hand-counted sample.
- Read the plan, not just the output. You confirm how many rows a query reads and why, then decide if it's good enough.
- Read-only on production. Agents explore production data through read-only access only, and customer data never leaves approved tools.
- Share what works. You turn useful prompts and checks into project skills or CLAUDE.md notes the team can reuse.
Nice to have
- Hands-on ClickHouse: AggregatingMergeTree, -State/-Merge combinators, argMax, windowFunnel, retention, dictionaries
- Python for analysis (pandas, Jupyter) alongside SQL
- Customer segmentation or clustering (k-means, RFM scoring) used in a live product
- Building dashboards or BI reports (Metabase, Superset, Looker or your own product UI)
- Domain knowledge in retail, restaurants, loyalty programmes or marketing campaign analytics
- Experience with campaign attribution: did a message lead to a visit, and over what window?
- Claude Code skills or prompts you've written for data work