Siddharth Singh

ABOUT

Siddharth Pratap Singh

AI/ML Enthusiast

Hi, I’m Siddharth Pratap Singh — a driven and curious self-taught developer, passionate about building intelligent systems that make a real impact. I specialize in Machine Learning, Deep Learning, and Full Stack Development, with a strong focus on AI-first applications.
Currently, I’m working on two flagship projects:

🤖 An AI Interview Feedback Bot that simulates real-time interviews with dynamic question generation and spoken feedback.
🦾 A Virtual Assistant for the Visually Impaired combining Computer Vision and Generative AI to describe real-world scenes in real time.

With a solid foundation in Python, NLP, Transformers, and deployment pipelines, I enjoy taking ideas from concept to production — building smart, useful, and scalable tools. I’m also actively exploring RAG (Retrieval-Augmented Generation), model fine-tuning, and LLM customization for domain-specific tasks. When I’m not coding, I’m learning, reflecting, or designing systems to help others succeed. I'm committed to becoming a versatile ML engineer who blends creativity with technical depth — and I’m just getting started.

Pyhton Machine Learning NLP Hugging Face

Download Resume

PROJECTS

NetFlix Dashboard

2025 • Data Analysis

1. A responsive dashboard for NetFlix dataset showcasing real-time analytics.

Data Analysis Power BI

SightSpeak

2025 • Deep Learning • Computer Vision

1. A computer vision-based assistant that uses real-time object detection
2. OCR, and scene captioning to describe the surroundings through audio
3. include sign language recognition and lip reading for the hearing impaired

Deep Learning OpenCV NLP TTS

AI Interview Feedback

2025 • NLP • Streamlit • Classic ML • RegEx

1. Implemented multimodal analysis: speech recognition (SpeechRecognition), sentiment scoring, and emotion detection
2. Role-Matching Model: Trained Linear SVM on 13K+ job descriptions (F1-score: 0.82)
3. Deployed secure containerized application with authentication (Docker)

LLMs NLP TTS Streamlit Docker

Customer Churn Prediction

2025 • Classic ML • GridVsearch • Hyperparameter Tuning • SHAP

1. Reduced incorrect churn alerts by 22% by tuning class-weighted XGBoost on imbalanced data
2. Optimized model with GridSearchCV and engineered 15+ features (e.g., usage trends, payment latency)
3. Used SHAP for feature impact explanation and created customer risk segment
4. Deployed interactive dashboard with Power BI and Streamlit Cloud

Machine Learning Model Optimizatin Streamlit Power BI

SKILLS

💻 Languages

Python

C++

JavaScript

📚 Libraries & Frameworks

Machine Learning

Deep Learning

Natural Language Processing

Voice & Audio

🧪 Development & Deployment

HTML/CSS/TailwindCss

Flask

Visualization & EDA

Git,GitHub

Streamlit

⚙️ Tools & Platforms

Google Colab / Jupyter Notebooks

Hugging Face Inference API

VS Code

Docker

CONTACT

Find me elsewhere

Latest Tweets

Virtual Avatar

🤖 An AI-powered representation of me — trained on my skills, experiences, and aspirations. Ask it anything, and it speaks for me.
- Made with Gemini 1.5 Flash
- Has a .json database to get info from
- Answers Questions Regarding my Portfolio