Software Engineer – GenAI

Job Title: Software Engineer – GenAI & Agentic SystemsDesigns, and develops advanced Generative AI applications with a specific focus on autonomous agents and multi-agent systems for our enterprise-grade network technology and autonomous platform. Leverages large language models (LLMs) to create reasoning engines that interact with external tools, APIs, and structured data. Specializes in building stateful, cyclic AI workflows using LangGraph to solve complex, non-linear problems in enterprise network.Translates high-level product requirements into robust Python code, bridging the gap between stochastic LLM outputs and deterministic software systems. Implements Retrieval-Augmented Generation (RAG), optimizes prompt engineering strategies, and deploys scalable agentic architectures that can visualize, plan, and execute multi-step tasks.What you will do Agentic Architecture Design: Design and implement stateful, multi-agent workflows using LangGraph. Define nodes, edges, and conditional logic to create agents capable of loops, self-correction, and human-in-the-loop interactions.GenAI Application Development: Develop end-to-end GenAI applications in Python, utilizing frameworks like LangChain and LlamaIndex to integrate LLMs (e.g., GPT-4, Claude, Llama 3) with enterprise data and APIs.Tool & Function Calling: Engineer robust "tools" and function-calling interfaces that allow agents to interact with internal databases, perform web searches, or execute code to complete tasks.Performance Optimization: Evaluate and improve agent performance. Move beyond basic accuracy metrics to measure "agent trajectory" quality, token usage efficiency, latency, and success rates in multi-step reasoning tasks.Collaboration & Integration: Collaborate with software engineers to expose agents via APIs (e.g., FastAPI) and integrate them into existing product ecosystems. Explain agent decision-making processes (observability) to non-technical stakeholders.Code Quality & Review: Contribute to design review sessions, specifically focusing on the maintainability of complex graph structures and state management. Provide feedback to peers on Python best practices and asynchronous programming.Prompt Engineering & Fine-Tuning: Design sophisticated system prompts and manage prompt versioning. Occasionally fine-tune Small Language Models (SLMs) for specific domain tasks where latency or cost is a constraint.Prototyping to Production: Rapidly prototype new agent concepts in stand-ups and move successful experiments to production-grade code, handling error recovery and guardrails against hallucination.What you need to bring Education and Experience RequiredBachelor's degree in Computer Science, Artificial Intelligence, Data Science, or a related quantitative field. Master’s degree preferred.Typically 2-4 years of total software or data engineering experience.Specific Requirement: At least 1 year of hands-on experience working with Generative AI, or building applications with LangChain/LangGraph.Knowledge and SkillsCore Technical Skills:Advanced Python: Strong proficiency in Python is non-negotiable, particularly with asynchronous programming (asyncio), Pydantic (for data validation), and typing.GenAI Frameworks: Deep practical knowledge of LangGraph (for graph-based flows) and LangChain. Familiarity with DSPy or AutoGen is a plus.LLM Orchestration: Understanding of Context Windows, Tokenization, Temperature, and the mechanics of Chat APIs (OpenAI, Anthropic, HuggingFace).Agentic Skills:Graph Theory & Logic: Ability to visualize and implement business logic as directed cyclic graphs (nodes, edges, conditional jumps).NoSQL Databases: Experience with Elasticsearch, redis or vector stores such as Pinecone, Milvus, Qdrant, or PGVector.Foundational Skills:Software Engineering: Proficient with Git, Docker, and CI/CD pipelines. Experience building and consuming RESTful APIs.Math & Stats: Solid understanding of linear algebra (embeddings) and probability as it relates to model sampling and decision-making.Communication: Ability to demystify "AI Magic" and explain the limitations (e.g., hallucinations) and capabilities of agentic systems to product managers and business stakeholders.