Applied AI Engineer

Note: All candidates meeting the basic screening threshold will complete a 30-minute timed online technical assessment tailored to the role (e.g., AI engineers get AI/LLM questions). This is a required step before interviews. Kindly ensure [email protected] is added to your safelist.

ABOUT THE ROLE

You will take AI features from the Product Lead's brief to production: design, build, evaluate, ship, and improve. You will work directly with the Product Lead, own the technical approach, and be expected to push back when a scope or design is wrong.

Our products are AI-native, multi-tenant SaaS platforms. Retrieval has to be accurate from day one, and every conversation should make the next release better. You are building systems that improve over time, not a static chatbot.

This role is for someone who has already shipped LLM features to real users. Developers whose AI experience is tutorials, notebooks, or demos will not meet the bar.

WHAT YOU WILL OWN

• RAG pipelines: Design and ship retrieval for multi-tenant applications, covering ingestion, chunking, embeddings, and re-ranking, with per-tenant data isolation
• LLM orchestration: Route requests across models on AWS Bedrock (Claude, Llama, and others) by task and cost, and build multi-step and agent workflows
• Evaluation: An automated benchmark with ground-truth questions that tracks retrieval accuracy and hallucination rate on a regular schedule
• Feedback and logging: Capture every conversation in a structured form (query, retrieved context, response, latency, user feedback) that works for evaluation and future fine-tuning
• Prompt and config control: Version every prompt template, RAG configuration, and chunking strategy with its measured results, so the team knows what was tried and what worked
• Production integration: Ship AI services behind clean APIs, working with backend engineers on reliability and cost
• Product judgment: Work with the Product Lead on what to build next and what not to build
• Mentoring: Guide junior engineers and fresh graduates on AI work through code review, pairing, and worked examples, and help raise the team's AI skills

MINIMUM

• 4+ years of software engineering experience
• 2+ years shipping LLM or RAG features to production
• Have shipped a RAG or LLM feature to real users and measured whether it worked, for example through retrieval accuracy or hallucination rate
• Can explain your chunking, embedding, retrieval, and evaluation choices, and why you made them
• Strong Python, and comfortable in TypeScript
• Have used a vector database in production
• Can own a problem from brief to production with little supervision

PREFERRED

• AWS Bedrock experience
• Multi-tenant system experience
• Fine-tuning workflows (Bedrock, SageMaker, LoRA) and RLHF concepts, even if you have not yet run training jobs
• Daily use of AI coding tools (e.g. Claude Code, Copilot)

STACK

Core: Python · TypeScript · AWS Bedrock (Claude, Llama) · a production vector database (Pinecone, Weaviate, or similar) · PostgreSQL · FastAPI · Docker

Evaluation: RAGAS, TruLens, LangSmith, or custom pipelines

Also useful: LangGraph, LangChain, LlamaIndex, Airflow, S3, Redshift

We care more about how you design and evaluate systems than how many tools you have used.

HOW TO APPLY

Your CV or portfolio must show, with concrete examples:

• AI systems you shipped to production: your exact role, the scale, and the outcome
• Design decisions and measurement: what you chose (chunking, retrieval, models), the alternatives you rejected and why, and how you measured quality
• Progression: how your scope and ownership have grown over your career

Applications without this evidence will not be reviewed. Please also include a GitHub or portfolio link and your expected monthly compensation. Shortlisted candidates complete the assessment above, and interviews will go deep on these examples.

Pay: Rs300,000.00 - Rs400,000.00 per month

Work Location: In person