Apply Below for a Direct HR Interview - First in India!

Job Title: AI Engineer – AI Compliance Platform

Employment Type: Full-time

Experience: 3–6 years preferred

Location: Ludhiana / Gurugram

Compensation: Open and negotiable based on experience and fit

About the Role

We are building an early-stage AI-powered product that helps businesses manage legal and regulatory compliance more effectively.

The first version of the platform is already live. Our next phase is focused on making its AI systems significantly more reliable, measurable and production-ready.

We are looking for an AI Engineer who can take ownership of the intelligence layer of the product—from evaluation and experimentation to production monitoring and continuous improvement.

A core part of this role is evaluation-driven development: defining what good looks like, building evaluation systems, identifying failure modes and using measurable signals to systematically improve AI performance.

This role is suited to someone who enjoys turning promising AI prototypes into dependable products that users can trust.

What You’ll Work On

• Design and maintain evaluation frameworks for AI-generated outputs
• Define quality metrics, acceptance criteria and benchmark datasets
• Build automated and human-reviewed evaluation pipelines
• Identify, categorise and analyse AI failure modes
• Use evaluation results to improve prompts, retrieval, workflows and model selection
• Design and improve agentic and multi-step LLM workflows
• Build reliable RAG pipelines for legal and compliance information
• Improve document ingestion, chunking, indexing, retrieval, reranking and grounding
• Develop citation, source-traceability and answer-verification mechanisms
• Design human-in-the-loop workflows for review, correction and escalation
• Implement guardrails for hallucinations, unsupported conclusions and unsafe outputs
• Monitor production AI performance, cost, latency and reliability
• Conduct model and prompt experiments using measurable comparisons
• Collaborate with Full Stack and UX/UI team members to turn AI capabilities into intuitive product experiences
• Contribute to technical and product decisions across the platform

What We’re Looking For

• Strong practical experience building applications using large language models
• Experience developing and evaluating RAG or retrieval-based systems
• Understanding of embeddings, vector search, reranking and information retrieval
• Experience designing prompts and structured LLM workflows
• Ability to define measurable quality standards for non-deterministic AI systems
• Proficiency in Python and experience working with APIs
• Familiarity with evaluation datasets, test cases and experiment tracking
• Understanding of hallucination, grounding, context management and model limitations
• Ability to diagnose AI failures and translate findings into systematic improvements
• Strong analytical thinking and attention to detail
• Comfort working in an early-stage environment where requirements may evolve
• A strong sense of ownership and accountability for production outcomes

Good to Have

• Experience with LLM evaluation frameworks or custom evaluation infrastructure
• Familiarity with agent frameworks and tool-calling systems
• Experience with hybrid search, metadata filtering and knowledge graphs
• Knowledge of document parsing, OCR and unstructured-data pipelines
• Experience building human feedback and review systems
• Familiarity with PostgreSQL and vector databases
• Experience with AI observability, tracing and production monitoring
• Understanding of model routing, caching, latency and cost optimisation
• Experience working with legal, regulatory or other high-accuracy domains
• Familiarity with cloud infrastructure, Docker and CI/CD pipelines
• Experience designing AI systems that handle sensitive or confidential information

How We Work

Our team believes in deep ownership with intentional overlap across AI, Full Stack and UX/UI.

You will own the AI layer deeply while thinking about the product as a whole. We value engineers who consider not only whether an AI workflow works in a demonstration, but also whether it is measurable, explainable, maintainable and dependable in real-world use.

We prefer evidence over intuition. Improvements should be supported by evaluations, experiments and production signals wherever possible.

What Success Looks Like

Within the first few months, you should be able to:

• Understand the existing AI architecture and its key failure modes
• Define clear quality metrics for the product’s important AI workflows
• Establish a repeatable evaluation and regression-testing process
• Build benchmark datasets representing real compliance use cases
• Improve the accuracy, grounding and consistency of AI-generated outputs
• Reduce unsupported answers and improve source traceability
• Strengthen the reliability of retrieval and agentic workflows
• Introduce effective production monitoring for quality, latency and cost
• Independently own AI improvements from hypothesis to production validation

Why Join Us?

• Work on a high-impact AI product addressing a real and complex business problem
• Solve challenging problems involving accuracy, trust and explainability
• Join at an early stage and directly influence the product and AI architecture
• Build production AI systems rather than isolated prototypes
• Take meaningful ownership of evaluation, reliability and continuous improvement
• Grow with the company as the product and AI team expand

Compensation

• Compensation is open and will be determined based on experience, technical ability and overall fit.