Head of Data and Analytics

About the Company

Adalat AI is building an end-to-end justice tech stack that automates manual and clerical pain points in courtrooms, giving judges back time to focus on what matters most: decision-making and delivering justice. Our solutions — from AI-powered transcription in Indian languages to case-flow management, document navigation, and the Paperless Courts platform — are now deployed across 10 states, covering nearly 25% of India's judiciary. Backed by leading technology companies and funders, and incubated at MIT and Oxford, Adalat AI is working to eliminate judicial delays and expand access to timely justice. Founded by a team with backgrounds in law, technology, and economics from Harvard, Oxford, MIT, and IIIT Hyderabad, we are scaling rapidly across India and the Global South.

Role Overview

We are hiring a Head of Data & Analytics to make this company's numbers trustworthy, and then to make them decisive.

We are deployed in courtrooms across 10 states. Every dictation, cause list, template and error is a data point, and we have years of them. What we do not yet have is a data function.

This is the role that fixes current analytics/data "fires" and then builds the thing that should have existed from the start: a defensible event taxonomy, instrumentation that engineering cannot bypass, a warehouse that other teams can query themselves, and a reporting layer that leadership, partnership teams and our government stakeholders plan around.

You will own the function end to end — the standards, the pipeline, the people, and the answers. You will work alongside the Product function, and your numbers and their evidence are expected to be checked against each other rather than published side by side.

We have a preference, not a requirement, for people who have built a data function from nothing rather than inherited a mature one. There is no data contract here, no naming convention, no verified event and no versioned mapping table. You will write the first version of all of them.

Key Responsibilities

Own the data contract — a naming convention, a verified-event register, and a review gate that no team ships instrumentation around
Fix the foundations — backfill and consolidate the properties the company already reports on, kill duplicate and free-text fields, and remove personally identifiable data sitting in analytics properties
Own the migration — the move to the new analytics platform and the warehousing solution
Set instrumentation requirements before schema changes ship — court establishment and hall context, versioned mapping tables with effective-from and effective-to dates, IDs rather than display labels as keys
Own the reporting layer — the bi-weekly feature report for leadership, state and feature reporting in the partnership portal, and the numbers that go to government stakeholders
Make the company self-serve — a modelled, documented warehouse other teams query themselves, so routine questions stop routing through your team
Lead the analysis that changes decisions — adoption and correlation work against partnership OKRs, and the investigations where the answer settles an argument
Build and run the team
Hold the line on what leaves the building — no number goes to a partner state or a CXO deck without a stated basis, a caveat where one is due, and a named owner

About You

You have cleaned up someone's analytics before and you are not romantic about it. You know the first ninety days are unglamorous — reading event definitions, finding out that two teams count the same thing differently, telling a CXO that a number in last quarter's deck was wrong. You are willing to do that work yourself rather than delegate it on day one.

You do not trust a number until you know how it was made. Your first question about a chart is what was filtered out. You have been caught by a silent join or a null-heavy property before, and it changed how you work.

You think in contracts, not dashboards. When engineering proposes a schema change, you can tell within a day which existing reports it breaks and what three properties need to ride along with it. You ask for those before the release, not after, and you ask in a way that does not block the release.

You can say no to senior people without making an enemy. Teams will want to ship events without review, pull numbers without context, and quote state-level figures you know are unsound. Holding that line politely, repeatedly, in writing, is a real part of this job.

You manage by raising the standard, not by taking the work. You review your analyst's SQL, you write the convention they follow, and you give them the visible deliverable rather than keeping it.

You write. Every number you publish arrives with a sentence saying what it means and a caveat you volunteered before anyone asked for it.

Qualifications

8+ years in analytics, data or business intelligence, with at least 2 years owning a function or leading analysts — title matters less than whether the standards were yours to set
Expert SQL — window functions, CTEs, incremental models; you review other people's queries and catch the join that silently changed the row count
Python or equivalent for reconciliation, cleanup and one-off analysis at a scale spreadsheets cannot hold
Deep hands-on ownership of a product analytics tool — Mixpanel, Amplitude, PostHog, GA4 or similar — at the schema level: events, properties, identity, lookup tables, not just the reporting UI
Warehouse and modelling experience — BigQuery, Snowflake, Redshift, Databricks or Azure — with a transformation layer such as dbt, and an opinion about how dimensions that change over time should be handled
You have designed an event taxonomy and an instrumentation review process that survived contact with a shipping engineering team
You have run at least one analytics platform migration or major cleanup end to end, and can describe what went wrong in it
You have presented numbers to executives and to external stakeholders, and have had to defend a methodology in the room
Clear, short writing — memos, definitions and caveats that non-technical people act on
Genuine interest in AI and legal-tech, and in how Indian courts take up technology

Especially valuable

Experience where analytics data is also training or evaluation data — ASR, ML or model quality correlation
Having built a metrics layer or semantic layer that other teams query without asking you
Government, judicial, healthcare or other regulated, privacy-sensitive environments, including handling PII inside analytics systems
Experience reporting to external or government partners on usage and outcomes
Using LLM tooling in the analytics workflow; we are scoping an analytics agent
Regional language ability — Telugu, Malayalam, Kannada or Hindi — given where we are deployed

What You Will Achieve in a Year

Every number this company quotes has a definition, an owner and a known level of confidence — and the ones that were quietly wrong have been found, corrected and communicated
The migration is complete and reconciled, the old stack is switched off, and nobody is maintaining two sources of truth
Instrumentation is gated: no product schema change ships without the analytics properties it needs, because you are in that conversation before the release, not after
State, court and establishment reporting is defensible at district level, not just High Court level, and partnership teams use it with government stakeholders without a caveat from you
Other teams query the warehouse themselves for routine questions, and your team spends its time on the questions that change decisions
You have a team of two or three with a clear bar, and at least one person visibly better at their job than when they joined

Our Interview Process

We keep our process straightforward and transparent. Here's what to expect:

R1 — Intro Call (30 minutes)
An introduction to Adalat AI — our mission, the problem we're solving, and an initial conversation around role fit.

R2 — Data-Analytics Deep-Dive (45 minutes)
A discussion on your past experience, problem solving process, and how your background connects to the work we're doing.

R3 — Project Review (60 minutes)
A walkthrough of your major project with a cross-functional panel — walk us through your work, decisions, and thinking.

R4 — Culture Fit — Founder Chat (30 minutes)
A conversation with one or more of our founders to assess mutual fit and shared values.

R5 — Offer
If it's a great match on both sides, we'll move forward with an offer.

Note: The process may vary depending on seniority and role. For lead roles, a whiteboarding challenge may be added after the portfolio review.