Backend Lead (Python / FastAPI)

Job Description:

About the Role

You will design and build the entire Evernine Brands data platform — the ingestion pipelines, schema normalisation engine, and API surface that the Flutter frontend and AI systems consume. You will work directly with the founder, alongside the Frontend Lead and a dedicated AI/ML Lead who owns the intelligence and prescription layer.

This is a pure data and API engineering role. You own the reliability, scalability, and correctness of the data platform. The AI/ML Lead takes clean, well-structured data from there and builds the intelligence layer on top.

Key Responsibilities

• Data ingestion pipelines: CSV parsing, validation, and normalisation across 13 ad platforms (Meta, Google, YouTube, LinkedIn, Flipkart, Truecaller, AdMob, and more) into a unified schema
• Unified schema engine: Core data model — platform | date | campaign | ad_id | creative_type | headline | body_copy | spend | ROAS | CTR | India+ fields — with intelligent field mapping and type coercion
• Cross-zone data joins: Linking paid ads data with owned content, social, and retail media signals using shared keys (ad_id, landing_page_url, asset_id) — producing clean, joined datasets for the AI/ML Lead to consume
• REST API: FastAPI endpoints consumed by the Flutter frontend — authentication, data sync, platform settings, and structured data delivery
• Background jobs: Async pipeline workers for CSV processing and URL crawling using Celery or ARQ
• Database architecture: PostgreSQL schema design, indexing strategy, and migration management with Alembic
• Infrastructure: Containerised deployment ensuring the data platform is observable, reliable, and scalable

How You Will Work With the AI Lead

The AI/ML Lead owns the intelligence layer — embeddings, vector search, LLM orchestration, and content prescriptions. Your job is to ensure they always have clean, well-structured, fully-joined data to work with. The interface between your roles is a set of well-defined Postgres tables and async job triggers. You will collaborate closely on schema design but own entirely separate layers.

Requirements

Must-have

• 5+ years of Python backend engineering in production systems
• Strong FastAPI or Django REST Framework experience — you have designed and shipped production APIs
• Deep PostgreSQL knowledge — schema design, indexing, query optimisation, and migrations with Alembic
• Experience building ETL or data ingestion pipelines — CSV/JSON processing, data validation, and transformation at scale
• Strong understanding of async Python — asyncio, background workers, and task queues (Celery, ARQ, or RQ)
• Production experience with Redis for caching and task queue brokering

Nice to Have

• Experience with web crawling and scraping (BeautifulSoup, Playwright, Scrapy) for content extraction
• Familiarity with ad platform data structures — Meta, Google Ads, or retail media CSV exports
• Knowledge of the Indian ad-tech ecosystem — programmatic buying, retail media, vernacular platform specifics
• Prior experience at a martech, adtech, or data-heavy B2B SaaS company
• Docker and container-based deployment — experience with Railway, Render, or AWS ECS

Core Tech Stack

Python 3.11+ · FastAPI · PostgreSQL · Redis · Celery / ARQ · Alembic · Docker

Architecture Context

The Evernine Brands backend has two layers you own entirely:

• Ingestion layer: CSV upload → validation → schema normalisation → Postgres. Each of 13 ad platforms has a defined field mapping. India-specific platforms (Truecaller, AdMob, gaming networks) carry non-standard fields (placement_type, completion_rate, add_to_cart_rate) requiring special handling.
• Joins layer: Cross-zone joins linking ad rows to their landing pages, creative assets, and social content using shared identifiers. This joined dataset is what the AI/ML Lead's pipeline reads from.

 The AI/ML Lead owns everything downstream of clean data: embeddings, vector search, LLM inference, and prescription output.