Computer Engineering student at Texas A&M building reliable AI systems: agentic workflows, LLM evaluation, retrieval-grounded applications, and full-stack developer tools.

Houston, TX · LinkedIn · Portfolio · aishasalimg@gmail.com

• Applied ML: multimodal classification, data pipelines, confidence calibration, model evaluation
• LLM systems: RAG, tool calling, structured outputs, AI-as-a-judge, reliability evaluation
• AI engineering: agent orchestration, human-in-the-loop escalation, workflow automation
• Software engineering: Python, TypeScript, .NET, React/Next.js, REST APIs, testing

Research project developing evaluation and reliability workflows for LLM-assisted RTL verification. Note: This work is currently private due to research and repository-access restrictions. I can discuss the system design, evaluation methodology, and my contributions upon request. Stack: Python, OpenAI API tool calling, SystemVerilog/UVM, Cadence Xcelium, JasperGold, JSON

Multimodal deep-learning system for skin-condition classification across skin tones. Highlights: 50K+ images and metadata · approximately 82–83% Top-3 accuracy
Stack: Python, TensorFlow/Keras, Xception, ResNet50V2, CNNs, pandas, NumPy

RAG-based p5.js learning assistant that grounds responses in course-specific materials. Focus: contextual retrieval, structured source grounding, and LLM-response reliability
Stack: Python, OpenAI API, RAG, prompt engineering

• Building evaluation and reliability workflows for tool-using LLM systems
• Contributing to open-source AI, observability, or developer-tooling ecosystems
• Seeking software engineering, ML engineering, and applied AI opportunities

AI/ML: TensorFlow/Keras, scikit-learn, pandas, NumPy, RAG, OpenAI API, Claude
AI Systems: Agentic workflows, MCP, tool calling, structured outputs, LLM evaluation
Engineering: Python, TypeScript/JavaScript, SQL, .NET, React, Next.js, Node.js, Git