Lead AI Engineer

Recrew AI; Delhi; On-site; INR 3,500,000 - 6,000,000; Full time

Benefits offered by this job

End-to-end IC ownership of AI products

Direct influence on product direction

Collaborative environment

Exposure to fast-moving AI stack

Job summary

Recrew AI in India is seeking a senior AI/ML engineer to own end-to-end production systems. You will transform fuzzy business asks into scalable architectures, pipelines, and measurable outcomes, with focus on speed, reliability, and cost.

You will mentor engineers, define design reviews, and own a slice of the AI product roadmap while shipping at scale for millions of users and advertisers.

Qualifications

• 6+ years of software engineering experience with real traffic and 3 a.m. failures.
• Proven track record shipping LLM-powered products in production.
• Hands-on experience with prompting, RAG pipelines, and LangGraph.
• Strong Python proficiency for AI/ML system development.
• Experience with AWS (or other cloud) including CI/CD and prod infra.
• Ability to translate business asks into technical designs independently.
• Strong written communication: specs, design docs, code.

Responsibilities

• Translate business-level asks into concrete technical designs — components, data flows, model choices, cost estimates, and build plan.
• Design full AI systems end-to-end: APIs, queues, storage, latency budgets, observability, and cost governance.
• Build and ship LLM-powered products using prompting, evals, guardrails, and multi-step workflows with state management.
• Architect and maintain RAG pipelines — embeddings, hybrid search, and index freshness.
• Measure offline evals, online A/B tests, and real numbers on quality, latency, and cost.
• Make model routing and hosting decisions across OpenAI, Anthropic, and open-source models.
• Mentor engineers, run design reviews, and own your slice of the AI product roadmap.

Skills

Python for AI/ML

Backend systems design

LLM product development

Production experience

System design

Tools

LangGraph

LangSmith

Braintrust

Retrieval systems (RAG)

Job description

Function: Artificial Intelligence / Machine Learning Engineering

Type: Full-time

About Company

One of India's largest private FM radio networks, reaching 40 million+ listeners weekly. The company broadcasts across 49 cities, 1,000+ towns, and 50,000 villages.

It is the only private FM station broadcasting from Jammu & Kashmir, with an international presence in Singapore and Bhutan. Its digital platform extends the brand into curated lifestyle, music, and news content.

A team of 740 operates at the intersection of legacy broadcast media and digital innovation. Engineers here build AI products that touch millions of listeners, creators, and advertisers daily.

Position Overview

The company is building AI products for listeners, creators, and advertisers — and this role sits at the centre of that effort. As a senior individual contributor, you will take fuzzy business conversations and turn them into shipped, measurable AI systems: designing the architecture, building the pipelines, and owning the outcomes. This is not a research role — the bar is production, not demos.

Role & Responsibilities

• Translate business-level asks ('smarter recommendations', 'automate the ad-copy thing') into concrete technical designs — components, data flows, model choices, cost estimates, and a build plan; write specs when needed and push back when AI is the wrong tool.
• Design full AI systems end-to-end: service boundaries, APIs, queues, caching, vector and SQL storage, latency budgets, failure modes, and observability dashboards — with cost-per-request and SLOs as first-class concerns.
• Build and ship LLM-powered products using structured prompting, few-shot patterns, prompt versioning, evals, and guardrails; implement multi-step agentic workflows using LangGraph or equivalent frameworks with state management, tool calling, and retries.
• Architect and maintain RAG pipelines — chunking strategies, embeddings, hybrid search and reranking, retrieval quality evaluation, and index freshness.
• Measure everything: offline evals, online A/B tests, and real numbers on quality, latency, and cost; 'it feels better' is a gut-check, not a launch criterion.
• Make model routing and hosting decisions across OpenAI, Anthropic, and open-source models; evaluate self-hosting trade-offs and fine-tuning ROI honestly.
• Mentor engineers, run design reviews, and own your slice of the AI product roadmap — keeping the team current on a fast-moving ecosystem without chasing every shiny repo.

Must Have Criteria

• 6+ years of software engineering experience with strong backend and system design depth — services that have met real traffic and real 3 a.m. failures.
• Demonstrated track record of building and shipping LLM-powered products in production — shipped, monitored, and costed; notebook demos do not count.
• Hands-on production experience with prompting (structured, few-shot, versioning, evals, guardrails), RAG pipelines, and agentic frameworks such as LangGraph.
• Strong proficiency in Python for AI/ML system development.
• Experience with AWS (primary) or another major cloud provider, including CI/CD pipelines and production-grade infrastructure.
• Demonstrated ability to take a business-level ask and independently produce the technical concept and design — more than once.
• Strong written communication: specs, design docs, and code that other engineers can read and build on.

Nice to Have

• Background in media, radio, streaming, or advertising technology.
• Experience with fine-tuning, eval tooling (Ragas, LangSmith, Braintrust, or equivalent), or ranking and recommender systems.
• Voice or audio AI experience — STT/TTS pipelines, audio understanding.
• Prior formal leadership experience — hiring, mentoring, growing a team; strong ICs looking to move in this direction are equally welcome.
• Startup experience with broad ownership across architecture, implementation, and operations.

What We Offer

• End-to-end IC ownership of AI products serving millions of listeners, creators, and advertisers on a greenfield media super-app.
• Direct influence on product direction — you shape what gets built, not just how.
• Collaborative environment with product and engineering teams working on high-impact personalisation, content intelligence, and advertising AI problems.
• Exposure to a fast-moving AI stack with the autonomy to make real architectural and tooling decisions.