Applied AI Research Lead

Logikality; Bengaluru; On-site; INR 1,500,000 - 2,100,000; Full time

Job summary

Logikality in Bengaluru is seeking a senior AI/ML engineer to lead production AI systems for document understanding in mortgage workflows. You will own model selection, evaluation harnesses and quality thresholds, driving reliable results across complex documents.

Proficiency in Python and modern AI tooling is essential, with a focus on prompting, retrieval and fine-tuning. Experience with regulated domains is a plus and mentoring engineers is welcomed.

Qualifications

• Experience shipping production AI/ML systems.
• Deep knowledge of LLMs, retrieval, evaluation and prompting patterns.
• Comfort defining metrics, running experiments and defending trade-offs with evidence.
• Strong software engineering foundation in Python and modern AI tooling.

Responsibilities

• Build production AI systems for document understanding and multi-document reasoning across mortgage and title workflows.
• Own model selection, evaluation harnesses and quality thresholds tied to real workflow outcomes.
• Evaluate performance with structured benchmarks, error analysis and human review sampling.
• Improve system reliability through prompting, retrieval, fine-tuning, verification and controlled fallbacks.
• Translate operator feedback into measurable model and system improvements.
• Lead technical decisions on architecture, tooling and applied research direction.

Skills

Production AI

LLM systems

Metrics & experiments

Python programming

Evidence-based decision making

Tools

Python

AI tooling

Job description

Mortgage operations depend on accuracy, evidence and defensible decisions. This role sets the technical direction for how our AI understands documents, reasons across a loan file and produces outputs that operators can trust and act on.

What you will own

• Build production AI systems for document understanding, extraction and multi-document reasoning across mortgage and title workflows.
• Own model selection, evaluation harnesses and quality thresholds tied to real workflow outcomes.
• Evaluate performance with structured benchmarks, error analysis and human review sampling.
• Improve system reliability through prompting, retrieval, fine-tuning, verification and controlled fallbacks.
• Translate operator feedback into measurable model and system improvements.
• Lead technical decisions on architecture, tooling and applied research direction.

Expected outcomes in the first 90 days

• Deliver a first evaluation harness for one production workflow with baseline metrics.
• Ship one measurable accuracy or reliability improvement to a live application.
• Establish a repeatable pattern for error analysis and regression tracking.
• Define the applied AI roadmap for the next two quarters with owner and outcome per initiative.

Experience and capabilities required

• Strong track record building and shipping production AI or ML systems.
• Deep working knowledge of LLM systems, retrieval, evaluation and prompting patterns.
• Comfortable defining metrics, running experiments and defending trade-offs with evidence.
• Strong software engineering foundation in Python and modern AI tooling.

Helpful additional experience

• Experience with document intelligence, OCR pipelines or structured extraction at scale.
• Prior exposure to regulated domains such as financial services, insurance or healthcare.
• Experience mentoring engineers or leading small technical teams.

Work model and location

In-office · Bangalore · Full-time

Hiring process

• Application review against the outcomes required for the role.
• Initial conversation on experience, interests and understanding.
• Role-relevant discussion, case or technical assessment.
• Final discussion on responsibilities, working model and expectations.