Technical Lead: AI Architect

Job title: Technical Lead_AI Architect

Department: Appalus Technology Practice (Full Stack, AI & Salesforce), deployed with a leading fintech client

Reports to: Technology Practice Director, Appalus; day-to-day alignment with the client's CTO / technology leadership

Location: Noida (hybrid), working from the client's office

Experience: 10-15 years total; 5+ years building AI/ML systems in production; 2+ years in BFSI / fintech preferred

Team: Leads 6-12 engineers across AI, data and platform within the client engagement; influence across the client's product squads

About the role:

Appalus is hiring a Technical Lead cum AI Architect, on Appalus payroll, to be deployed with one of our key clients: a leading Indian fintech that processes very high volumes of digital financial transactions every day through a nationwide network.

This is a decision-making leadership role, not an API-integration role. Working as an extension of the client's technology leadership, you will set the AI architecture and governance framework, choose models and vendors, lead the engineering team that builds AI features, and advise senior stakeholders on where AI creates value and where it creates risk.

You will work where regulated financial systems meet applied AI: fraud and risk signals, customer and partner service automation, operations intelligence, developer productivity, and internal knowledge systems, all under RBI, DPDP Act 2023 and PCI DSS obligations. You will also be part of the Appalus Technology practice, with access to our AI delivery community, certifications and Claude partner ecosystem.

Key responsibilities:

AI strategy and architecture:

• Define the client's AI reference architecture: LLM gateway, RAG and agent patterns, model routing, observability, and fallback paths for high-throughput, low-latency transaction systems.
• Own build-vs-buy and model selection decisions (Claude, GPT-4o, open-source models) on cost, latency, accuracy, data residency and vendor risk.
• Build a prioritized AI use-case roadmap with business owners (fraud and risk, KYC/onboarding, collections, customer support, reconciliation, engineering productivity), each with a measurable business case.
• Design architectures that keep AI out of the critical payment path unless it meets strict latency, availability and audit standards.

AI governance, risk and compliance:

• Establish and chair the AI governance framework: AI policy, model inventory, risk tiering of use cases, approval gates, and a model risk management process aligned with RBI guidance.
• Ensure DPDP Act 2023, PCI DSS and data-localization compliance across AI workloads: PII masking, consent, retention, and no leakage of customer or card data to third-party models.
• Define AI quality and safety standards: automated evals, red-teaming, prompt-injection defence, hallucination detection, bias and fairness checks, and explainability for decisions affecting customers.
• Put human-in-the-loop controls, audit trails and incident response in place for every production AI system.
• Lead vendor and tool assessments (Claude, Cursor, GitHub Copilot, internal LLM deployments), including security reviews and contract terms on data usage.

Technical leadership and delivery:

• Lead and grow a team of AI, data and backend engineers; run design reviews and code reviews for ML and LLM code.
• Own the end-to-end lifecycle of AI features: prototype, evaluate, deploy, monitor, retire.
• Oversee data pipelines for training, evaluation and monitoring (Airflow, dbt, Spark, Snowflake) with lineage and data-quality controls.
• Set MLOps / LLMOps standards: versioning, CI/CD for models and prompts, drift monitoring, cost tracking, and SLAs.

Organization wide AI adoption:

• Act as the trusted AI advisor to the client's CXO team and keep Appalus leadership aligned on engagement outcomes: explain trade-offs, risks and ROI in plain business language.
• Drive AI enablement across teams through playbooks, training, internal communities and rollout of AI copilots for engineering.
• Mentor engineers on applied AI, evaluation practice and responsible AI usage.
• Track and report AI adoption, value delivered and risk posture to leadership every quarter.

Required qualifications:

• 10-15 years of software engineering, including 5+ years designing and running AI/ML systems in production and 3+ years leading engineering teams.
• Architect-level experience with high-volume, low-latency distributed systems (event streaming, microservices, high availability), ideally in payments, lending, banking or other transaction-heavy domains.
• Deep proficiency in Python; working knowledge of TypeScript, Java or Go.
• Hands-on experience with LLM frameworks (LangChain, LangGraph, LlamaIndex) and agentic architectures.
• Proven delivery of RAG systems, vector databases (pgvector, Pinecone, Weaviate) and semantic search pipelines.
• Strong grasp of model evaluation: offline evals, human feedback loops, A/B testing, hallucination detection, and latency/cost benchmarking.
• Production experience with cloud AI services (AWS Bedrock/SageMaker, Azure ML or GCP Vertex AI), Docker, Kubernetes and Terraform.
• Practical experience setting up AI governance or model risk controls and working knowledge of data privacy and security regulation (DPDP Act, PCI DSS, RBI IT and outsourcing guidelines, or equivalents such as GDPR).
• Familiarity with data engineering and ETL pipelines feeding ML workflows.
• Excellent communication: able to make and defend architecture decisions with CXOs, auditors, regulators and engineers.

Preferred qualifications:

• Fintech AI use cases in production: fraud detection, credit scoring, transaction categorization, AML/KYC automation, or conversational banking.
• Fine-tuning or RLHF on open-source models (LLaMA, Mistral, Qwen or similar) and self-hosted LLM deployment for data-residency needs.
• Background in NLP, information retrieval, anomaly detection or recommendation systems.
• Familiarity with responsible AI frameworks such as NIST AI RMF, ISO/IEC 42001 or the EU AI Act.
• Experience rolling out AI copilot tooling (Cursor, GitHub Copilot, Claude) at team or org level and measuring productivity impact.
• Cloud architect or ML certifications (AWS, Azure or GCP); security certifications are a plus.

AI and engineering stack:

Languages: Python (primary), TypeScript, Java/Go, SQL

LLM / AI: Anthropic Claude, OpenAI GPT-4o, open-source LLMs, LangChain, LangGraph, LlamaIndex

Vector DBs: pgvector, Pinecone, Weaviate

Infra: AWS (Bedrock, SageMaker), Docker, Kubernetes, Terraform

Data: Kafka, dbt, Airflow, Snowflake, Spark

Governance and observability: LLM gateway, eval harness, prompt/model registry, PII redaction, audit logging

Dev tools: Cursor (enterprise), GitHub Copilot, Claude Team