Senior Software Engineer - Agentic AI - Python Expert

Speria; Atlanta (GA); On-site; USD 180,000 - 230,000; Full time

Job summary

Speria is seeking senior engineers to build fast, AI-first agentic systems. You will design, ship, and operate multi-agent AI services that integrate LLMs with tools, planning, memory, and evaluation.

The role emphasizes production-grade Python backend, plus integration with our enterprise stack (TypeScript/Angular, .NET/C#, SQL Server, Azure). You will work on memory/grounding, MCP-based tool access, and agent-to-agent communication, delivering low-latency experiences in real customer

Qualifications

• Proven experience building LLM-powered applications with Azure OpenAI, embeddings, vector stores, RAG, prompt engineering, and evaluation pipelines.
• Hands-on with agent frameworks such as Semantic Kernel, LangGraph, LangChain Agents, AutoGen, or CrewAI.
• Ability to design deterministic, evaluatable, and safe agent behaviors including function schemas, tool success metrics, fallback strategies.
• Practical use of Prompt Flow for authoring, testing, and deploying multi-step AI workflows in Azure AI Foundry.

Responsibilities

• Build agentic AI applications on Azure AI Foundry, using Azure OpenAI models, prompt flow, tool-calling and evaluation.
• Design memory architectures (episodic/semantic/vector/graph) and RAG pipelines to improve factuality and latency.
• Implement MCP services to standardize tool access across agents and enable agent-to-agent communication.
• Connect agents to Microsoft Fabric and Dataverse workflows, ensuring governance and auditability.
• Develop AI-native backend services in Python (FastAPI) with observability and cost/latency dashboards.
• Embed AI features into Speria stack (TypeScript/Angular, .NET/C#, SQL Server) and CI/CD pipelines.
• Use AI-native development tools (GitHub Copilot, Bolt, Cursor, Replit, vibe-coding) to accelerate delivery.
• Enforce safety and reliability: guardrails, red-teaming, PII protection, prompt hardening, regression tests.
• Own the full cycle: design → model/tool selection → API/UI → deployment → monitoring → continuous improvement.

Skills

Azure OpenAI

LLM apps

Agent frameworks

Python backend

Evaluation pipelines

Tools

Semantic Kernel

LangGraph

LangChain Agents

AutoGen

CrewAI

Job description

We are hiring senior engineers who build fast, think AI-first, and can take agentic AI from prototype to production. You will design, ship, and operate agentic systems that combine large language models (LLMs), tools/functions, planning, memory, evaluation, and multi-agent communication. You will work primarily in Python for AI services and integrate with our enterprise stack (TypeScript/Angular, .NET/C#, SQL Server, Azure), delivering trustworthy, cost-efficient, low-latency experiences in real customer workflows.

What You'll Do!

• Build agentic AI applications on Azure AI Foundry: Azure OpenAI models, Prompt Flow, tools/function-calling, evaluations, vector search (Azure AI/Cognitive Search), and orchestration for multi-step reasoning and tool use.
• Design memory & grounding: implement episodic/semantic/long-term memory with vector/graph stores; architect RAG pipelines and retrieval strategies that improve factuality and reduce latency/cost.
• Integrate via Model Context Protocol (MCP) to standardize tool/skill access; design agent-to-agent communication, delegation, and event-driven workflows.
• Connect agents to Microsoft Fabric (OneLake, Lakehouse, Warehouse, Real-Time Analytics) and Dataverse entities/workflows; ensure lineage, governance, and auditability.
• Develop AI-native backend services in Python (FastAPI, asyncio) with evaluation harnesses, observability, and cost/latency/quality dashboards.
• Embed AI features into the Speria stack: TypeScript/Angular UIs, .NET/C# services, SQL Server, NServiceBus, Azure DevOps pipelines, and Ionic/Cypress where applicable.
• Use AI-augmented development tools like GitHub Copilot, Bolt, Cursor, Replit, and vibe-coding workflows to accelerate delivery, test generation, refactoring, and documentation.
• Implement safety & reliability: guardrails, red-teaming, PII protection, prompt hardening, regression tests, automated evaluations; uphold SLO/SLA excellence in production.
• Implement full cycle agentic engineering: design → model/tool selection → API & UI → deployment → monitoring → continuous improvement.

What You Bring!

Core AI & Agentic Expertise

• Proven experience building LLM-powered applications with Azure OpenAI, embeddings, vector stores, RAG, prompt engineering, and evaluation pipelines.
• Hands-on with agent frameworks such as Semantic Kernel, LangGraph, LangChain Agents, AutoGen, or CrewAI.
• Ability to design deterministic, evaluatable, and safe agent behaviors including function schemas, tool success metrics, fallback strategies.
• Practical use of Prompt Flow for authoring, testing, and deploying multi-step AI workflows in Azure AI Foundry.

MCP, Memory & Agentic Communication

• Experience building and consuming MCP services to standardize tool access across agents.
• Implemented memory architectures (episodic, semantic, vector, graph) and long-running conversational context.
• Designed agent-to-agent communication patterns (messaging, orchestration, delegation, arbitration).

Microsoft Data & App Platform

• Integration with Microsoft Fabric, SQL Server, Supabase, Databricks (OneLake/Lakehouse/Warehouse/Real-Time) for grounding data, retrieval, and telemetry.
• Working knowledge of Dataverse entities, actions, and triggers; connecting agents to line-of-business records and Power Platform workflows.
• Databricks for ELT, Delta Lake pipelines, feature engineering, ML training/serving, MLflow tracking and model lifecycle.
• Azure IoT Hub/IoT Edge pipelines to incorporate device telemetry and edge-to-cloud intelligence into agentic workflows.
• Azure services: App Service/Functions/AKS, Key Vault, Storage, Event Hubs/Service Bus, Monitor/Application Insights.

Python & Backend Engineering

• Production-grade Python (FastAPI, asyncio, type hints), Postgres/SQL, Redis, queues, OpenTelemetry, CI/CD, and containerization.
• Strong API design, testing (unit/integration/property-based), performance tuning, and reliability engineering.

Front-End & Speria Enterprise Stack

• Experience in TypeScript/Angular for operator consoles and human-in-the-loop oversight.
• Ability to integrate with .NET/C#, SQL Server, NServiceBus and Azure DevOps in our enterprise environment.

AI-Native Dev Workflow & Culture

• Daily use of GitHub Copilot, Bolt, Cursor, Replit, and vibe-coding to speed delivery and raise quality.
• Mentor teams in prompting, agent behavior design, context management, evaluation, and AI-assisted engineering practices.
• Seasoned aptitude for action, tight feedback loops, crisp written communication, and ownership mindset.

Success Looks Like (Outcomes)

• Quality & reliability: rising agent tool-use success rate; falling hallucination/retry rates; low incident volume; fast MTTR.
• Performance & cost: P50/P95 latency and token-cost budgets met; measurable efficiency gains across services.
• Adoption & impact: shipped features used by real users; clear business KPIs improved via automation/intelligence.
• Engineering excellence: high test coverage, stable CI/CD, observable systems, and healthy on-call posture.

Tooling & Stack Summary

• AI & Agentic: Azure AI Foundry (Azure OpenAI, Prompt Flow, evaluations), MCP, Semantic Kernel, LangGraph, LangChain, AutoGen, CrewAI, HuggingFace embeddings, vector DBs, Azure AI/Cognitive Search, RAG, memory architectures.
• Data & Integration: Databricks (ELT, ML, Delta Lake, MLflow), Microsoft Fabric (OneLake/Lakehouse/Warehouse/Real-Time), Dataverse, Event Hubs/Service Bus.
• IoT: Azure IoT Hub, Azure IoT Edge, stream ingestion & device telemetry flows.
• Services: Python (FastAPI, asyncio), .NET/C#, REST/gRPC, containers, CI/CD with Azure DevOps.
• Frontend: TypeScript/Angular, Ionic; E2E testing with Cypress.
• AI-Native Dev Tools: GitHub Copilot, Bolt, Cursor, Replit, vibe-coding workflows.