- Advanced Python proficiency: Mastery of Python syntax, data structures, OOP principles, and AI/NLP libraries (LangChain, LangGraph, Transformers, OpenAI SDK, HuggingFace, LlamaIndex).

- Web frameworks: Hands-on expertise with FastAPI, Flask, or Django for backend service development; ability to lead projects and mentor developers.

- API development leadership: Design and development of secure, high-performance RESTful APIs, with strong understanding of authentication mechanisms and integration patterns.

- Databases & ORM: In-depth experience with relational databases (PostgreSQL, MySQL, SQLite) and ORM frameworks (SQLAlchemy, Django ORM).

- Microservices & Docker: Architecting modular, scalable, resilient microservices-based systems with containerization.

- CI/CD & DevOps: GitHub workflows, Jenkins, automated build/test/deploy pipelines.

- Testing & quality: TDD discipline with PyTest/unittest; robust focus on code quality, maintainability, and reproducibility.

- Algorithms & problem-solving: Design, analysis, and optimization of algorithms; strategic problem-solving for complex challenges.

AI & Agentic Engineering Expertise

- Hands-on LLM experience: Building applications with foundation models (OpenAI, Claude, LLaMA, Mistral, Gemini) via APIs and self-hosted/open-weight models.

- Agentic AI systems: Designing autonomous/semi-autonomous agents with multi-step reasoning, tool/function calling, planning, memory, and state management — using LangGraph, LangChain, or similar frameworks (e.g., AutoGen, CrewAI, Semantic Kernel, Model Context Protocol (MCP) for tool integration).

- Prompt engineering & RAG: Structured prompts,



few-shot and chain-of-thought techniques, embedding models, vector stores (FAISS, Chroma, Weaviate, pgvector), hybrid search, chunking strategies, and re-ranking for retrieval pipelines.

- LLMOps & evaluation: Systematic evaluation of LLM outputs — golden datasets, LLM-as-judge, regression testing of prompts/chains; observability and tracing of agent workflows (e.g., LangSmith, Langfuse, OpenTelemetry for GenAI); handling hallucination mitigation, fallback mechanisms, and graceful degradation.

- Fine-tuning & model adaptation: Familiarity with parameter-efficient fine-tuning (LoRA/PEFT), instruction tuning, and when to choose fine-tuning vs. RAG vs. prompting.

- Structured outputs & tool integration: Function calling, JSON-schema-constrained generation, and integrating LLMs securely with APIs, databases, and monitoring platforms.

- MLOps & inference infrastructure: Containerized AI workloads, GPU orchestration, batch vs. streaming inference, model serving (e.g., vLLM, Triton), token/cost optimization, caching, and rate limiting.

- AI security & governance: Data privacy, PII redaction, prompt-injection defenses, guardrails frameworks, role-based access, auditability,



and responsible/ethical deployment of LLM systems at enterprise scale.

- Multi-modal & emerging capabilities: Experience with multi-modal models (text logs/metrics/diagrams), streaming responses, and voice/chat interface design.

In this role, you will:

- Collaborate closely with Product Owners, Network SMEs, and Engineers to gather requirements and identify AI opportunities in telemetry analysis, event correlation, configuration analysis, and operational summarization.

- Own end-to-end development of network tooling services across telemetry, topology, configuration management, and orchestration — and augment them with GenAI capabilities.

- Architect and deploy agent-based systems that integrate with network management platforms and decision-support systems.

- Design and implement prompt flows, vector-based search, and retrieval pipelines to contextualize large-scale telemetry and logs.

- Build natural-language interfaces enabling intelligent self-service for operations teams (querying, reasoning, diagnostics, root cause analysis).

- Design, implement, and publish high-performance APIs serving Maersk's network operations ecosystem.

- Drive architectural decisions; mentor engineers on software engineering best practices and on LLM usage, architecture, evaluation, and security in production.

- Support program delivery with estimation, planning, and risk assessment; participate in capacity planning, performance optimization, and reliability initiatives.

- Continuously evaluate emerging GenAI models, agent frameworks, and tools for Ops Intelligence use cases

-

📌 Senior Software Engineer (Pune)
🏢 A.P. Moller - Maersk
📍 Pune