Open for Opportunities

Full Stack Developer with 3+ years of experience across the entire engineering spectrum - Python & Node.js backends, React & Vue.js frontends, data pipelines, DevOps, and production-grade agentic AI systems. I architect and ship complete platforms from the ground up, combining deep software engineering roots with cutting-edge AI Platform Engineering.

2+ Years in AI
70+ Agents Shipped
500+ Enterprise Users
73 MCP Tools Built
mario_bridge.py
?
🏆 You earned Mario's respect, champ!
mario_bridge.py Edit and click ▶ Run
1
2
3
4
5
6
$ Ready? Write your code and click ▶ Run

I build AI that works
in production.

I started my career as a Full Stack Developer building scalable backends with Python and Node.js, and polished frontends with React.js, Vue.js, and Vanilla JavaScript. Over time I evolved into AI Engineering focusing on Data & Scalable Platforms, making me a complete engineer who can own any layer of the stack.

Today I specialise in AI Platform Engineering designing and shipping agentic systems, MCP servers, RAG pipelines, and multi-agent runtimes. I open-sourced MCP servers with 73 tools and 15 resources in SAP and authored AgentOS, a multi-agent runtime that delivered 85% productivity improvement and 90% reduction in onboarding time.

I care about clean architecture, measurable impact, and building AI systems that don't break when they hit real users.

Full Stack Development

Python & Node.js backends, React.js & Vue.js frontends, REST & GraphQL APIs full ownership from DB to UI.

AI & Agentic Engineering

LangGraph, MCP, LiteLLM, Pydantic AI - multi-agent runtimes and RAG pipelines in production.

Data Engineering

Data pipelines, vector databases (Qdrant, Pinecone), semantic retrieval and hybrid search systems.

DevOps & Cloud

Docker, Kubernetes, Terraform, CI/CD - infrastructure validated at 10x load spikes.

Where I've Made an Impact

AI Engineer - AI Platform Engineering (Working Student)

SAP Signavio  ·  Sankt. Leon Rot / Heidelberg

September 2025 - March 2026
  • First engineer to build and open-source MCP servers at SAP Signavio 73 tools, 15 resources, now the foundation for all LLM integrations.
  • Shipped 5 production AI agents (RCAA, Conversational Diagnostics, Insights Voice, BPMN, Process Intelligence) serving 500+ enterprise users, cutting analyst time by 30%.
  • Designed LiteLLM routing infrastructure enabling multi-provider LLM calls (Claude, OpenAI, Gemini) with unified fallback and cost controls.
  • Contributed 4 modules to the PINTS Process Intelligence Agent (GraphQL/MCP binding, Pydantic AI validation, LangChain workflows) strengthening the MAS backbone orchestrating AI agents across SAP
  • Collaborated with GTM and senior engineers on BTP infrastructure (Terraform, Docker, Kubernetes HPA/VPA) and resolved 5 post-launch production incidents with no SLA breaches
PythonLangGraphMCPRAG FastAPIReactKubernetesLiteLLM

Full Stack Developer - SAP CPIT (Working Student)

SAP SE  ·  Walldorf / Heidelberg

March 2025 – September 2025
  • Built the Strategic Portfolio Roadmaps App end-to-end (React, Node.js, TypeScript, Python) with an automated PPT generator that processes Jira data into roadmaps adopted by 700+ stakeholders, saving 70% of planning cycle time
  • Owned full delivery across 3 sprint cycles: UI5 frontend, Python/python-pptx layout engine, and investment cluster grouping logic shipped in close collaboration with the Dev Manager and Product Owner
  • Collaborated with S/4 HANA SoftwareArchitects and Staff Software Engineers to Build data pipelines for Demand Deliuvery Dashboards(DDT) using ABAP CDS, RAP, S/4 HANA Backend and other ABAP Techniques (CDC, CAP)
SAPUI5React.jsNode.jsTypeScriptABAPPython FlaskKubernetesJiraSAP BTP

Master Thesis - AgentOS

SRH University Heidelberg

2025 – 2026
  • Designed AgentOS - A multi-agent orchestration runtime for enterprise process intelligence workflows.
  • Achieved 85% productivity improvement and 90% reduction in manual analyst onboarding time.
  • Implemented dynamic agent routing, structured output validation (Pydantic AI), and tool-use orchestration across specialised sub-agents.
LangGraphPydantic AIFastAPI DockerMulti-Agent

Software Engineer

Diebold Nixdorf  ·  Hyderabad

March 2022 - May 2023
  • Monitored 1500+ ATM software server endpoint nodes with Grafana alerting, caught 3 critical failures before user impact and maintained 99.7% ATM network uptime across the production fleet and spike outage
  • Led RCA for a high-severity ATM outage affecting 50,000+ daily transactions restored service in 4 hours and implemented safeguards with zero recurrence over the next 9 months
  • Automated ATM and ERP provisioning with Ansible playbooks cut setup time from 3 hours to 25 minutes per server across 40+ instances, saving 80 onsite technician/engineer-hours per quarter.
LinuxNetworkingIncident ResponseAnsibleGrafanaPrometheusShell ScriptingBash

Things where I have made Impact

5 Production ReAct AI Agents

SAP · 500+ Users

Shipped 5 production AI agents at SAP Signavio - RCAA (Root Cause Analysis), Conversational Diagnostics, Insights Voice, BPMN Agent, and Process Intelligence - cutting analyst time by 30%.

LangGraphRAGQdrantLiteLLMGraphQL

Kubernetes Autoscaling Backend

10x Load Validated

Designed and validated a Kubernetes-based autoscaling backend for AI agent workloads using HPA and custom metrics. Infrastructure as Code via Terraform. Validated at 10x load spikes in production.

KubernetesTerraformDockerFastAPIHPA

Stock Market Dashboard

100% Live Market Insights

Conversational Stock Live Dashboard – Fullstack AI Microservices Application This project is a Live stock market dashboard built with a microservices architecture. It enables users to interact with stock data in real-time using websockets that fetch real time stock data using 3rd party API's.

FastAPIJava Spring BootReact.jsGraph entitiesRedisQdrantNeo4jWebsocketsGemini-Flash-001RAG PipelinesMicroservicesReal-time Data

Purchace Order to Pay

100% Live Market Insights

AI to create and organize new requests for procurement. If users want to buy a product or service they need to create a formal request to the procurement department. That will afterwards process this request. But with This Agent it retrieves all data from the PO and creates a new request for procurement uses redis for same bytes matching. this practice prevents faulty rejex parcing and AI Model calling.

FastAPIBytesIOOpenAIRedishashlib

What I Work With

Frontend Development

React.js90%
Vue.js85%
JavaScript / Vanilla JS95%
TypeScript80%
HTML / CSS95%

Backend Development

Python95%
FastAPI / Flask90%
Node.js / Express85%
GraphQL / REST APIs90%
PostgreSQL / MongoDB75%
Git / GitHub95%

AI & Agentic Engineering

LangGraph / Agentic AI95%
MCP (Model Context Protocol)95%
RAG Pipelines (Qdrant, Pinecone)90%
LiteLLM / Claude / OpenAI / Gemini90%
Pydantic AI85%
Prompt Engineering95%

DevOps & Cloud

Docker88%
Kubernetes80%
Terraform75%
CI/CD Pipelines78%
Linux82%
GCP / AWS72%

Let's get in touch
You'll hear back from me within - 24 hours.

I'm open to senior engineering roles, AI platform projects, and research collaborations. Based in Heidelberg, Germany - available for remote and hybrid opportunities.

Send a Message
Mohith Tummala Full Stack Developer · AI Platform Engineering Heidelberg, Germany