AI Researcher · Cognitive Intelligence · AI Safety
Designing transparent, safe, and cognitively grounded reasoning systems.

I work at the intersection of cognitive science, artificial intelligence, and AI safety, with an emphasis on building systems that can reason, explain, and align with human values.

• 🧠 Cognitive Intelligence Systems (ACT-R / SOAR-inspired reasoning)
• 🧩 AI Safety & Interpretability (chain-of-thought monitoring, alignment)
• ⚙️ High-Performance AI Systems (CUDA, scalable learning & simulation)
• 🤝 Human–AI Reasoning Transparency

A non-clinical cognitive AI system designed to:

• Infer intent, emotion, and cognitive load
• Reason using cognitive architectures
• Provide transparent recommendations (music, reading, reflection)

🔗 Repository: work in progress
📄 Research notes: in preparation

A technical AI safety framework to:

• Inspect internal reasoning traces of LLMs
• Detect hallucinations and unsafe reasoning paths
• Improve interpretability and auditability

🔗 Repository: in preparation

• Cognitive AI & Reasoning Systems
• Technical AI Safety & Alignment
• Mechanistic Interpretability
• High-Performance Computing for AI
• Human–AI Interaction & Transparency
• AI for Mental Health & Education

Languages & Frameworks
Python · PyTorch · TensorFlow · Hugging Face · CUDA

Systems & Libraries
Linux · OpenCV · NumPy · Pandas · Scikit-learn

• 🔗 LinkedIn: https://linkedin.com/in/sumit45
• 📧 Email: [email protected]

Building AI systems that reason, explain, and align.