Tier-1 India (Bangalore, Hyderabad, Chennai, Pune, Mumbai, Delhi NCR) · Remote-friendly · Contractor · Reports to the CTO

About the client

Our client is a Singapore-headquartered, AI-native product carbon footprint and ESG intelligence platform serving automotive OEMs and their tier 1, 2 and 3 suppliers, with engineering operations in India. Their current build gives suppliers an automated, screening-grade carbon footprint through public-data ingestion, AI-powered gap-filling and a confidence-tiered pipeline, with no analyst in the loop.

About the role The Data Analyst is the primary resource who marshals the documentation that builds the platform's reference data layer. The platform produces a credible screening-grade carbon footprint by ontological equivalence: estimating a supplier's missing data from comparable suppliers matched on material, process and country. That only works once a reference base of roughly 1,500 suppliers populates the equivalence space.

Building that base is what makes meaningful output possible for any customer, so this role is foundational and sits on the critical path.

At launch the platform serves EU OEMs and their tier-1 suppliers, whose supplier base is global. The reference base you build is therefore global and multi-language, not India-only.

You will not do this alone or by hand. You work AI-first, using Claude to extract, translate and structure across languages, running on the pipeline the AI/ML Platform Engineer builds. The base is built coverage-first: EU and India, the high-yield and high-relevance regions, come first for launch, then China, Southeast Asia and the Americas as expansion. You run the ingestion. You do not build it.

Source and marshal the documentation, globally

- Global sources: EU company registers and CSRD-cascade disclosures, national registries, OEM and tier-1 supplier portals, Indian filings (BRSR, MCA), Japanese and Korean filings, industry associations such as CLEPA, VDA, ACMA and SIAM, and global databases, all per the approved-source registry.
- Multi-language and AI-first: use Claude to extract,



translate and structure non-English sources at scale. You curate and QA. The AI does the volume.
- IP discipline: ingest derived attributes only and never redistribute source documents. Every field carries a source and a confidence tier of A, B or C.

Build the base, coverage-first

- EU and India first: build coverage of the equivalence neighbourhoods (material, process, country) for EU OEM and tier-1 suppliers in the high-yield regions first, so launch can ride on a covered sub-footprint.
- Default-anchor the hard regions: where public data is sparse, as in China, Southeast Asia and North Africa, anchor with regional and sector defaults plus a few representative suppliers rather than fully sourcing each, then improve as primary data arrives.
- Structure and resolve: normalise into the defined schema, resolve entity-resolution and source-format exceptions, and maintain one canonical profile per supplier.

Quality, provenance and drift

- QA against the golden set: check profiles within tolerance, triage and flag gap-fills, and keep provenance and confidence tier on every field.
- Monitor drift: watch source-format changes across regions and keep the source registry current.

What success looks like Progress is measured by supplier coverage, not by time in seat.

- Foundation: taxonomy, ingest contract, entity ID and quality gates frozen. Per-region automated yield measured. Golden set seeded.
- Early cohorts: cohort model proven, with EU and India yield confirmed on real cohorts.
- Core coverage: EU and India equivalence neighbourhoods started, producing the first meaningful screening outputs.
- Launch-ready: EU and India covered, so an EU OEM or tier-1 gets credible screening results for its EU and India suppliers.



This is the usefulness threshold for general availability.
- Post-launch expansion: extend coverage to China, Southeast Asia and the Americas, multi-language and default-anchored where data is sparse.

Required qualifications

- Data literacy: spreadsheets and SQL, with some Python a plus. You can structure messy source data into a defined schema.
- Research and sourcing skill across jurisdictions: finding and extracting the right data from reports, filings and portals, with strong attention to detail.
- AI-first working: fluent using Claude and similar tools to extract, translate and structure at scale, applying human judgement on exceptions and QA.
- Comfort with multi-language sources via AI translation, and rigour on provenance covering source, date and confidence tier.
- 2 to 4 years in a data, research, analyst or operations role. High-volume data collection and QA experience is a solid plus.
- Willingness to learn the carbon accounting, LCA and automotive supply-chain domain. Prior domain knowledge is not required.

Boundaries This is a scoped role.

- You own the sourcing, structuring and QA of the global reference base data.
- Not this role: building the ingestion pipeline (AI/ML Platform Engineer), taxonomy and methodology governance (CTO), the programme (Programme Manager), or the customer-facing product (Product Engineer).
- Seam: you operate the pipeline the AI/ML Platform Engineer builds, escalate taxonomy questions to the CTO, and take benchmark tolerances from the LCA Analyst.

Team and engagement terms

- Team: likely a lead analyst working AI-first, scaling to a small pool for the global expansion.
- Type: contractor, with global sourcing scope.
- Window: full-time from August to December during the reference-base build, moving to fractional from January to April.
- Location: Tier-1 Indian cities preferred (Bangalore, Hyderabad, Chennai, Pune, Mumbai, Delhi NCR), remote-friendly.
- Reports to: the CTO.
- IP: NDA and IP-assignment agreement required before the engagement commences.

📌 Data Analyst (Contract) (India)
🏢 Colleve
📍 India