About Experience Skills Projects Awards Blog ↗ Contact
Dublin, Ireland

Vinith
Kumar G P

AI Engineer · Agentic AI Systems

I build AI systems that actually make it to production. Multi-agent orchestration with LangGraph, RAG pipelines, and tool-using agents on AWS, with 5+ years of Python engineering behind it all.

Download Resume View My Work Read My Notes
Find me on
5+
Years Experience
10M+
Daily Transactions Processed
85%+
PyTest Coverage Sustained
MSc
Data Analytics (ML & AI), NCI

// about me

An AI engineer who ships.

I'm an AI Engineer based in Dublin, Ireland, with 5+ years of experience spanning Python engineering, data science, and production AI systems that handle real scale.

At Infidata Technologies, I built AI-powered workflows for a financial platform processing 10M+ daily transactions. That meant designing agent-like reasoning loops, RAG retrieval architectures, and GPT-based API integrations on AWS. Along the way I cut release cycles by 40% and kept test coverage above 85% with PyTest and GitHub Actions.

I then did my MSc in Data Analytics (Machine Learning & AI) at NCI Dublin, where my capstone was a production-ready multi-agent AI platform built with LangGraph orchestration, MCP integrations, and an LLM-as-judge evaluation harness.

These days I'm at Nando's Ireland as an Operations & Performance Analyst, running forecasting and KPI analysis across 5 live dashboards and using AI tooling to speed up the work. I've also mentored 100+ engineers and interns and taught Python to 500+ students across 10+ colleges.

  • Dublin, Ireland
  • vinithvkk050@gmail.com
  • MSc Data Analytics (ML & AI), National College of Ireland
  • Open to AI/ML, agentic AI, and Python engineering roles
⚡
10M+
Daily transactions handled by my AI decisioning pipeline
🚀
40%
Faster release cycles via CI/CD automation
🧪
85%+
PyTest coverage sustained in production
🏆
1st
Place, NCI Hackathon (LITTLE LIFE)

// experience

Where I've shipped.

Production AI systems, real scale, real teams.

Operations & Performance Analyst
May 2024 – Present
Nando's Ireland · Dublin
  • Monitor and analyse 5 operational dashboards daily covering sales, labour, cost, and customer metrics, tracking 10+ KPIs to spot trends and flag anomalies before they escalate.
  • Identify 3 to 4 recurring performance gaps per quarter through trend and variance analysis, with recommendations that lead to real efficiency improvements.
  • Run weekly sales forecasting and actuals-vs-targets analysis across multiple revenue streams, using AI tools and automation to accelerate analytical workflows.
  • Translate complex data patterns into concise, stakeholder-ready insights that directly influence operational and resourcing decisions.
Data AnalysisForecastingKPI Tracking AI-Assisted WorkflowsStakeholder Communication
Senior Python & AI Engineer
Jan 2021 – Dec 2022
Infidata Technologies · India
  • Delivered AI-powered workflows for a financial platform processing 10M+ daily transactions, designing agent-like reasoning loops and custom Python orchestration pipelines with GPT-based APIs and Hugging Face models on AWS.
  • Made production AI features more reliable by building scenario-based evaluation frameworks and RAG-style retrieval architectures for multi-step AI workflows.
  • Cut release cycle time by 40% while keeping test coverage above 85%, using automated PyTest suites and GitHub Actions CI/CD to take AI services from proof-of-concept to production.
  • Applied pragmatic engineering judgement on model selection, prompting, and orchestration per use case; presented prototypes in architecture reviews with product and data science teams.
  • Mentored 100+ interns and junior engineers through code reviews, workshops, and pair programming.
PythonGPT APIsHugging Face RAGAWSFastAPI PyTestGitHub Actions
Junior Python Engineer
Jul 2019 – Dec 2020
Infidata Technologies · India
  • Maintained reliable REST API endpoints and Python data ingestion pipelines within Agile squads on PostgreSQL and AWS, with automated and manual test coverage.
  • Reduced critical query execution time by 35% by optimising SQL queries and indexing strategies on high-volume PostgreSQL databases.
  • Supported deployment and monitoring of production backend services on AWS, picking up observability, alerting, and structured logging fundamentals along the way.
PythonSQLPostgreSQL AWSREST APIs
MSc in Data Analytics (Machine Learning & AI)
2023 – 2024
National College of Ireland · Dublin
  • Specialised in Machine Learning & AI, Applied Predictive Analytics, Cloud Computing, Database Design, and Data Mining & Statistical Analysis.
  • Won 1st place in the NCI Hackathon with LITTLE LIFE, a real-time maternal and fetal health risk prediction system built using ML.
  • Capstone: production-ready multi-agent AI orchestration platform with LangGraph, MCP, RAG, and a comprehensive evaluation harness.
Machine LearningPredictive Analytics Cloud ComputingStatistical Analysis

// skills

What I work with.

No made-up percentages. Just the tools I actually use in production.

🤖
Agentic AI & LLM Engineering
LangGraphLangChain MCPRAG Pipelines Prompt EngineeringLLM Evaluation OpenAI APIClaude API Hugging Face
⚙️
Python & Backend
Python (Advanced)FastAPI FlaskDjango REST API DesignAsync Programming MicroservicesOOP
☁️
Cloud, MLOps & DevOps
AWS LambdaAPI Gateway S3 · DynamoDBCloudWatch AWS CDKDocker GitHub Actions CI/CD
🗄️
Databases
PostgreSQLMySQL MongoDBDynamoDB SQL (Advanced)Query Optimisation
🧠
Data Science & ML
scikit-learnXGBoost TensorFlowKeras Pandas · NumPyStatistical Analysis Power BI
🧪
Testing & Practices
PyTest (85%+ coverage)Unit & Integration Testing LLM-as-Judge EvalsAgile / Scrum Code ReviewJIRA

// projects

Things I've built.

Production AI systems with real metrics, not just side projects.

🛡️
Featured · Agentic AI
SupportIQ
A support agent that knows when not to answer. A LangGraph agent retrieves from the knowledge base, grounds every reply in sources, then decides to answer, hold for a human, or escalate. Refunds, cancellations, and data-deletion requests always pause on a real checkpointed interrupt until a person approves. Benchmarked against a naive answer-everything bot on the same 33 questions.
100%
Routing Accuracy
0 / 7
Unsafe Auto-Sends (naive bot: 7/7)
100%
Grounded Answers
PythonLangGraphLangChain RAGOpenAIFlask PyTest
🧪
Featured · AI + AutoML
AutoDS Copilot
Point it at a spreadsheet and it runs the whole ML workflow: explores the data, flags leakage and silent data problems, detects the task, compares dozens of models on a cross-validated leaderboard, tunes the winner, then writes the analysis in plain language. Same engine exposed three ways, a web console, a LangGraph agent with human-in-the-loop, and an MCP server any AI client can call.
36
Models Benchmarked
3
Interfaces (Web · Agent · MCP)
0
Data Leakage by Design
PythonLangGraphMCP RAGscikit-learnXGBoost LightGBMFlask
🛰️
Featured · Open Source
SKU Sentinel
Statistical governance for demand forecasts, so 50,000 SKUs don't need 50,000 sign-offs. One Flask app with an exception queue ranked by revenue at risk, an 8-model forecasting bench (seasonal naive to XGBoost to a PyTorch LSTM), causal promo-uplift analysis, and a drift monitor that grades its own detection quality. Validated on real Walmart data (M5) across three forecaster regimes.
86%
Drift Recall
83%
Alarm Precision
12/12
Data Breaks Caught
PythonFlaskscikit-learn XGBoostLightGBMPyTorch SciPyPlotly
💳
🏢 Production · Infidata
High-Volume Fraud Decisioning Engine
Real-time AI decisioning service on AWS Lambda processing 10M+ daily transactions for a financial platform. Cut operational overhead by 50% and false positives by 25%, with PyTest suites wired into GitHub Actions CI/CD so releases went out without surprises.
PythonFastAPIAWS Lambda DynamoDBDockerPyTest
🏆
🥇 1st Place · NCI Hackathon
LITTLE LIFE
Maternal and fetal health risk prediction system built at NCI Dublin's hackathon. An ML classifier that predicts risk in advance from CTG data. Cost-effective, real-time, and it took first place out of all competing teams.
PythonMachine Learning scikit-learnHealthcare AI
🛒
Customer Propensity Model
ML propensity model that predicts which website visitors will convert into paying customers, and shows which factors actually drive the purchase decision.
Pythonscikit-learn Predictive Analytics
📊
IT Incident Log Analyser
4-page Power BI dashboard turning ServiceNow audit-system incident logs into resolution insights, with a dynamic CRM layer via JetAdmin for managing incident records.
Power BISQL JetAdminData Visualisation

// accomplishments

Beyond the code.

Awards, publications, and community work.

🏆
1st Place · NCI Dublin Hackathon
Won first place at National College of Ireland's hackathon with LITTLE LIFE, an ML-based maternal and fetal health risk prediction system.
👨‍🏫
Mentored 100+ Engineers & Interns
Led Python development mentoring at Infidata Technologies through code reviews, workshops, and pair programming. Project completion rates went up 45%.
🎓
Python Workshop Instructor · 500+ Students
Delivered Python development workshops across 10+ colleges in India, teaching REST API design, backend architecture, and Python best practices.
⭐
Best Outgoing Student · ECE Department
Awarded Best Outgoing Student of the Electronics & Communication Engineering Department, Sai Vidya Institute of Technology, Karnataka.
📄
Published Research · SQL Injection & Prevention
Presented paper on "Recent research on SQL Injection attacks and prevention" at the 3rd National Conference on Recent Trends in Electronics and Communications (2017), Bengaluru.
🌐
International Conference · Prague, Czech Republic
Presented research at the 7th International Conference on Photonics, Optics and Laser Technology in Prague. Public speaking at an international level, early in my career.

// contact

Let's work together.

I'm currently open to AI/ML engineering, agentic AI, and Python roles in Dublin and across Ireland.
If you're hiring, or know someone who is, reach out directly.

vinithvkk050@gmail.com linkedin.com/in/vinith-kumar-g-p GitHub Blog & Notes
Download Resume Send an Email