Applied AI Student · Rosenheim, Germany

Hi, I'm Abhi Faldu.
I build industrial AI systems.

From anomaly detection in machine sensor data through real-time IIoT monitoring to production-ready MLOps pipelines. I turn messy signals into models that ship — with Python, PyTorch, MLflow and Docker.

abhi.py

class Engineer:
    name      = "Abhi Faldu"
    role      = "AI / ML"
    focus     = ["MLOps", "IIoT",
                 "Anomaly Detection"]
    stack     = ["PyTorch", "MLflow",
                 "FastAPI", "Docker"]
    location  = "Rosenheim, DE"

    def status(self):
        return "Open to AI & Data roles"
4Production ML projects
0.94Best F1-score (bearing PdM)
90%Test coverage on MLOps pipeline
6thSemester, B.Sc. Applied AI

01 — About

Turning sensor noise into decisions

I'm an Applied Artificial Intelligence student at Technische Hochschule Rosenheim, currently in my 6th semester. My hands-on work spans the full lifecycle of industrial AI — from anomaly detection in machine sensor data, through real-time IIoT monitoring, to production-ready MLOps pipelines.

I like problems where the data is messy, the constraints are real, and the model has to actually run in production. I care about experiment tracking, drift monitoring, tests, and clean deployment — not just a notebook that scores well once.

Before Germany, I worked as a Junior Data Analyst in India, building Power BI dashboards and EDA pipelines. I'm ready to contribute practical solutions in an AI or data science team from day one.

02 — Skills

Tools I build with

ML & Modelling

  • Scikit-learn
  • XGBoost
  • PyTorch
  • TensorFlow
  • NLP
  • Predictive Modelling

MLOps & Deployment

  • MLflow
  • Docker
  • FastAPI
  • CI/CD (GitHub)
  • Evidently
  • MQTT

Data & Visualisation

  • Pandas
  • NumPy
  • Power BI
  • Grafana
  • Matplotlib
  • Jupyter

Programming & Tools

  • Python
  • SQL
  • GitHub
  • Streamlit
  • InfluxDB

03 — Projects

Selected work

SensorsAI4I 10k TrainMLflow CI GatePR-AUC .83 ServeFastAPI MonitorEvidently

MLOps Pipeline for Predictive Maintenance

06/2026 — Present

Production-grade MLOps workflow for a machine-failure classifier (AI4I 2020, ~10,000 sensor records): training, experiment tracking and a model registry in MLflow, with automatic champion promotion.

  • Raised PR-AUC with XGBoost from 0.38 baseline to 0.83; CI enforces PR-AUC ≥ 0.75 before every promotion.
  • Deployed via FastAPI, data-drift monitoring with Evidently, containerised with Docker Compose — 38 tests, ~90% coverage.
  • MLflow
  • XGBoost
  • FastAPI
  • Evidently
  • Docker
View on GitHub →
3 Robotstemp/vib/A MQTT1 Hz Z-scoreanomaly InfluxDBstorage Grafanadashboard

IIoT Simulator for Smart-Factory Monitoring

05/2026

An Industry 4.0 simulation: 3 robot arms stream sensor data (temperature, vibration, motor current) over MQTT every second, with real-time anomaly detection via a sliding-window Z-score.

  • Storage in InfluxDB, visualised through Grafana and a custom web dashboard.
  • Orchestrated as 6 services with Docker Compose.
  • MQTT
  • InfluxDB
  • Grafana
  • Docker
View on GitHub →
VibrationNASA IMS LSTM AEPyTorch AnomalyF1 0.94 DashboardStreamlit

Predictive Maintenance for Robotic Bearings

03/2026 — 05/2026

An LSTM autoencoder (PyTorch) for anomaly detection in bearing vibration data (NASA IMS Bearing dataset), catching failures early.

  • F1-score 0.94 for early detection of bearing failures.
  • FastAPI backend + Streamlit dashboard for live anomaly scores and maintenance recommendations — 3 services via Docker Compose.
  • PyTorch
  • LSTM
  • Streamlit
  • FastAPI
View on GitHub →
OHLCV17.5k hrs FeaturesRSI/MACD Modelwalk-fwd CV RiskATR stop TestnetBinance

ML-Based Crypto Trading System

12/2025 — 03/2026

A modular end-to-end pipeline for algorithmic BTC/USDT trading: ~17,500 hourly data points, 16 engineered features (RSI, MACD, ATR).

  • Walk-forward cross-validation to prevent leakage; ATR-based stop-loss and drawdown protection.
  • Deployed on Binance Testnet — honestly documented 50% out-of-sample accuracy: public market features carried no real predictive edge.
  • Python
  • Pandas
  • Backtesting
  • Risk Mgmt
View on GitHub →

04 — Experience & Education

The path so far

Sales Associate — Lidl GmbH & Co. KG

Present

Prien am Chiemsee, Germany

Reliable work in a high-volume retail environment while studying full-time — strong time management and resilience — while building professional German and intercultural teamwork.

B.Sc. Applied Artificial Intelligence

10/2023 — Present

Technische Hochschule Rosenheim, Germany

6th semester. Focus: Machine Learning, Deep Learning, Neural Networks, Data Science, Database Systems, Software Engineering, IT Security.

Junior Data Analyst — MR Infoware

04/2023 — 10/2023

Rajkot, India

End-to-end EDA and data-preprocessing pipelines, Power BI dashboards for KPIs, and support on ML model evaluation — translating business requirements into analytical insight.

B.Tech Information Technology (partial)

01/2021 — 07/2022

Atmiya Institute of Technology & Science, Rajkot, India

Completed two semesters of foundational engineering studies.

05 — Contact

Let's build something

I'm open to Werkstudent, Praktikum and entry-level AI / Data roles. The fastest way to reach me is email.