Staff AI Engineer

This role is critical to building the core intelligence engine that will power Practo's next-generation healthcare experiences. This is not an API-wrapping exercise; we are building scalable, high-performance AI infrastructure from the ground up. We need a foundational builder who can operate with high autonomy and set the engineering standard for AI-native development.
Requirements:
Deep Backend Core and Infrastructure Chops (~8 Years Experience). They are, first and foremost, a hardcore software engineer. Crucially, we are not looking for a pure Machine Learning Engineer or Data Scientist who spends their time training models or working exclusively in Jupyter notebooks. They must have spent years building, scaling, and maintaining complex backend systems. They know what architectural decisions cause bottlenecks at scale, how to design robust APIs, and how to build systems that don't break under heavy load.
The Pure Builder (100% IC): They are a dedicated, high-impact Individual Contributor. Their leverage comes from their technical architecture, the scalability of their systems, and their raw code output. They lead by technical example and architectural vision.
The AI-Augmented Developer: " They don't just build for AI; they build with AI. They are fluent in the new paradigm of software engineering, aggressively leveraging tools like Claude Code, GitHub Copilot, and Cursor to multiply their output. They can build end-to-end applications from scratch at a velocity that wasn't possible two years ago.
Plugged into the AI Ecosystem. They possess a deep, practical understanding of the modern AI software stack. They go beyond surface-level API calls and understand vector databases, LLM orchestration frameworks, retrieval-augmented generation (RAG) pipelines, and model serving infrastructure. They know how to integrate AI components into a broader, scalable software architecture.
High Agency and Startup Hustle" They operate with extreme resourcefulness. In a high-ambiguity environment, they do not wait for perfectly scoped requirements or external blockers to clear. They have a growth mindset, figure out the path forward, and ship relentlessly.
How to Find Them:
This is a highly competitive profile. We are looking for a rare intersection of rigorous traditional backend scaling experience and cutting-edge AI fluency.
Where to Source (Target Pools):
AI-Native Startups: Engineers currently at Series A-C startups building foundational AI tools, developer tools, or AI-first enterprise SaaS.
Platform/Core Teams at Tech Tier 1/2 Look for teams labelled "Core Infrastructure, " "Backend Platform, " or "Applied AI Platforms" at companies known for heavy engineering cultures (e. g., Swiggy, Razorpay, Zepto, Flipkart, or the India core-engineering hubs of global tech majors).
Keywords and Resume Signals:
Backend/Scale (Primary Signal): Distributed Systems, Microservices, Go, Rust, Java/Kotlin, Kubernetes, Kafka, gRPC, High-throughput, Low-latency.
AI/ML Integration Stack: Vector Databases (Pinecone, Weaviate, Milvus), RAG, LLMOps, Model Serving / Inference (Triton, vLLM).
Action Verbs: Look for Architected, Built from scratch, Designed the platform, Scaled backend from X to Y.
Green Flags to Look For in Screening:
Side Projects/Hacking: They actively build their own AI products or integrations on nights and weekends just to test new models or frameworks.
Tooling Obsession: When asked about their workflow, they enthusiastically detail how they use Claude, Cursor, or Copilot to automate boilerplate and write tests.
Pragmatism: They can articulate when not to use an LLM or complex AI solution, showing they value engineering pragmatism over hype.
Anti-Patterns (Red Flags for this specific role):
The "Pure ML" Modeller: Candidates whose experience is heavily dominated by model training, feature engineering, data science, and hyperparameter tuning, but lack a proven track record of building and deploying high-throughput backend services and APIs.
The "Prompt Engineer": Profiles that are heavy on prompt tuning and building basic UI wrappers/chatbots, but lack the 8+ years of deep backend/database/infrastructure scaling experience.
The Architect-Only: Candidates who draw diagrams but haven't pushed production code themselves in the last 12-18 months. We need someone who is hands-on.
Over-indexing on Process: Candidates who heavily emphasise Agile ceremonies, Jira management, and team sizes over technical outcomes and architecture.

Experience
5-9 yrs