AVP - AI Engineer - Pune Location (Full Time Opportunity)

Experience from 5 to 10 Years

Role : AVP - AI Engineer

Location : Pune

Shift Timings : 12:30 PM - 9:00 PM

AI and Data Engineer, Agentic AI Platform

We are building a next-generation enterprise AI and data platform that enables teams to design, deploy, integrate, and scale intelligent applications and agents across investment research, portfolio management, risk, client, and operational workflows.

The platform provides standardized infrastructure, reliable data products, commercial and self-hosted model access, retrieval, agent orchestration, evaluation, observability, security, and governance. It enables teams to rapidly deliver AI-powered capabilities in a scalable, secure, and production-ready manner.

We are looking for hands-on AI and Data Engineers who can:

• Build reliable, reusable, and AI-ready data products for investment and enterprise use cases.

• Design and develop production-grade AI applications, agents, and intelligent workflows using Python.

• Implement agent orchestration, tool integration, retrieval, memory, and multi-step workflows.

• Integrate AI capabilities into enterprise applications, data platforms, and business systems.

• Contribute to the shared AI and data platform that powers these capabilities at scale.

This role operates at the intersection of AI engineering, data engineering, agentic AI, platform engineering, and investment technology. It emphasizes Python, enterprise integration, reliable data products, security, CI/CD, observability, model and agent evaluation, and rigorous validation of production AI systems.

 What We’re Looking For:

Core AI and Data Engineering

• 5+ years of experience in software, data, AI, or platform engineering.

• Strong production development experience with Python and SQL.

• Proven ability to build and operate AI applications, data pipelines, APIs, data products, and distributed systems.

• Experience integrating AI capabilities into enterprise applications and business workflows.

• Strong understanding of:

• Object-oriented and asynchronous Python

• APIs, microservices, and application integration

• Data modeling, contracts, quality, reconciliation, metadata, and lineage

• Batch, streaming, event-driven, and API-based architectures

• Workflow orchestration, distributed processing, and production operations

• Experience with cloud data and AI platforms such as Azure, AWS, Google Cloud, Snowflake.

• Ability to design scalable, reusable, resilient, and operationally supportable services.

Generative AI and Agentic Systems

• Hands-on experience with:

• Large language models, embeddings, prompt engineering, context engineering, and RAG architecture.

• Commercial and open-source model APIs and ecosystems.

• AI agents, agentic workflows, reusable skills, tool calling, structured outputs, and multi-step orchestration.

• Vector, keyword, metadata-filtered, graph-based, and hybrid retrieval.

• Agent frameworks such as LangGraph, Microsoft Agent Framework, Semantic Kernel, AutoGen, Pydantic AI, or equivalent platforms.

Preferred experience includes:

• MCP servers and clients, agent-to-agent communication, and enterprise search.

• Semantic layers, ontologies, and knowledge graphs.

• Agent memory, state management, and human-in-the-loop controls.

• Document intelligence and multimodal processing.

• Reusable agent tools, skills, manifests, and registration standards.

Model Hosting and Inference

• Experience with one or more of the following:

• Azure OpenAI, OpenAI APIs, Anthropic Claude, Google Gemini, or Vertex AI.

• Hugging Face, vLLM, Ollama, Qwen, or other approved open-source model technologies.

• Containerized model deployment and GPU-based inference.

• Model serving, API gateways, quantization, memory optimization, and model configuration.

• Context-window, token, concurrency, batching, throughput, and caching optimization.

• Benchmarking and monitoring model quality, latency, availability, resource utilization, and cost.

Platform Engineering, Security, and DevOps

Strong experience with:

• CI/CD, infrastructure as code, automated provisioning, and governed deployment.

• API-first and microservices architectures.

• Docker, Kubernetes, AKS, GKE, or equivalent container platforms.

• Authentication, authorization, secrets management, OAuth, managed identities, RBAC, and Microsoft Entra ID or equivalent technologies.

• Repository governance, branch protection, pull-request validation, dependency controls, and software supply-chain security.

• Metrics, distributed tracing, structured logging, monitoring, and alerting.

• Designing observable, resilient, secure, and production-ready AI systems.

• Experience using AI-assisted development tools such as Claude Code, GitHub Copilot, Cursor, Codex, or comparable tools for:

• Code generation, refactoring, debugging, testing, and documentation.

• Data-pipeline, infrastructure, and CI/CD development.

• Codebase understanding, modernization, architecture extraction, and pull-request review.

• Governed development supported by automated testing, peer review, security scanning, and controlled deployment.

Candidates must be able to independently validate AI-generated code for:

• Functional correctness and test coverage

• Security and data-access risks

• Performance and scalability

• Maintainability and architectural alignment

• Dependency, licensing, and software supply-chain concerns

• Compliance with enterprise engineering standards

Evaluation, Observability, and Testing: Required

• Hands-on experience with data-quality, model-evaluation, LLM-evaluation, or AI-observability frameworks.

Experience building:

• Model benchmarking and evaluation pipelines

• Data-quality and reconciliation tests

• Retrieval and RAG evaluations

• Agent test harnesses and scenario-based simulations

• Continuous validation and regression testing

• Production monitoring for data, retrieval, model, and agent quality

Ability to measure:

• Data accuracy, completeness, freshness, and consistency

• Retrieval relevance and coverage

• Groundedness, factual consistency, citations, and provenance

• Tool selection, execution, and structured-output compliance

• Agent workflow success, failure, and escalation behavior

• Latency, throughput, reliability, resource utilization, and cost

• Familiarity with OpenTelemetry, OpenInference, Datadog, Arize, LangSmith, MLflow, Microsoft Foundry evaluation capabilities, or equivalent tools.

• Ability to use evaluation results to inform model selection, routing, release, rollback, and production-readiness decisions.