I am a Machine Learning Engineer focused on building practical AI systems that move from research ideas to real-world deployment.

My projects are pretty diverse like computer vision, object detection, OCR, facial landmark detection, LLM-based applications, multi-agent systems, synthetic data generation, and production ML infrastructure.

I enjoy building complete AI pipelines: from data collection and annotation, to model training and evaluation, to deployment on cloud, backend systems, and edge/mobile devices.

Currently, I am pursuing an M.Sc. in Artificial Intelligence at Brandenburg University of Technology Cottbus-Senftenberg, Germany and want to contribute to research advancing ML models for which I am working on gaining and advancing relevant skills. I also enjoy reading and testing out new innovations.

• Computer Vision and Object Detection
• fine-tuning and edge deployment
• OCR pipelines using Tesseract
• Facial Landmark Detection
• Synthetic Data for AI training
• LLM applications with LangChain, LangGraph, and LlamaIndex
• Multi-agent AI systems
• Retrieval-Augmented Generation
• FastAPI-based AI backends
• MLOps, CI/CD, Docker, and cloud deployment
• Explainable AI for biomedical machine learning

• Fine-tuned YOLOv5 for custom Chilean license plate detection.
• Built a self-annotated dataset from phone-captured vehicle images.
• Optimized the pipeline for offline mobile inference.
• Achieved approximately 5 FPS on-device.
• Deployed using TensorFlow.js.
• Integrated Tesseract OCR for real-time license plate text extraction.

Tech: YOLOv5, TensorFlow.js, Tesseract OCR, Computer Vision, Edge AI, Object Detection

• Replicated the Fake It Till You Make It synthetic-data approach.
• Used Microsoft FaceSynthetics dataset.
• Trained a facial landmark detection model from scratch using synthetic data only.
• Implemented label adaptation for landmark alignment.
• Studied synthetic-to-real generalization for face analysis.

Tech: PyTorch, Computer Vision, Facial Landmark Detection, Synthetic Data, Deep Learning

• Built a cross-modal deep learning system that generates speech from facial images.
• Combined computer vision with neural speech synthesis.
• Designed a pipeline for facial feature extraction and audio generation.

Tech: PyTorch, Multimodal AI, Computer Vision, Speech Synthesis

• Developed deep learning models for drug sensitivity prediction.
• Used gene expression datasets from cancer cell lines.
• Applied Particle Swarm Optimization for genomic feature selection.
• Used SHAP to identify interpretable oncogene biomarkers.

Tech: Deep Learning, Explainable AI, SHAP, PSO, Biomedical AI

• Designed backend infrastructure for an AI-powered financial analytics platform.
• Built multi-agent workflows that generate SQL and Python analysis pipelines.
• Created autonomous dashboard generation from AI-generated insights.
• Used Redis and MongoDB for low-latency retrieval and deterministic dashboard reconstruction.
• Deployed production services using FastAPI, Docker, WebSockets, Temporal, and Azure CI/CD.

Tech: FastAPI, LangChain, Multi-Agent Systems, SQL, Python, Redis, MongoDB, Docker, Azure

I worked on BioMarkerX, a deep learning model for drug sensitivity prediction using CCLE and GDSC gene expression datasets.

The project used:

• Deep learning for biomedical prediction
• Particle Swarm Optimization for feature selection
• SHAP for explainable AI
• Gene expression data for biomarker discovery

📄 Publication:
Identification of Biological Markers in Cancer Disease using Explainable Artificial Intelligence

Here are some of my public repositories and areas of work:

• Movie Streaming Application with Recommendation System in C# and Firebase
• Cosine Similarity using TF-IDF with OpenMP, MPI, and CUDA
• Online Shopping Management System using C# and MySQL
• Graph Visualizer using NetworkX in Python
• Boolean Retrieval Model for Proximity and Boolean Queries

I am currently focused on:

• Production-ready AI systems
• Computer vision deployment on mobile and edge devices
• LLM agents for data analytics
• Multimodal deep learning
• Synthetic data for model training
• Explainable AI and interpretable machine learning
• Building stronger open-source AI projects

• English: C1
• German: A2

I am open to opportunities in:

• Machine Learning Engineering
• Computer Vision Engineering
• Applied AI Engineering
• LLM Engineering
• AI Backend / MLOps Engineering
• Research Engineering

Building practical AI systems from data to deployment.