Amendra singh

I’m a Computer Science student specializing in Reinforcement Learning from Human Feedback (RLHF) and Supervised Fine-Tuning (SFT). I’m driven by building high-integrity systems that evaluate LLM reasoning, detect hallucinations, and translate complex algorithmic, statistical, and mathematical edge cases into precise, structured data payloads.

I’ve developed data infrastructure and evaluation pipelines with a strong emphasis on validation, latency, and reliability. In my projects, I built an automated data cleaning & EDA pipeline that achieved a “0.00% error rate” on “150K+ entries,” and I created a statistical evaluation sandbox that reduced evaluation latency by “40%.”

I also contribute to open-source AI workflows, especially around statistical data validation. My contributions include optimized data-cleaning scripts, refactoring legacy outlier detection for “25%” faster execution, and reviewing “30+” pull requests for technical accuracy and documentation clarity.

Recognized by Microsoft with a “Machine Learning Contribution Award,” I bring an analytical, performance-driven mindset to every validation and audit task. I enjoy turning research ideas—like Spiking Neural Networks (SNNs) and STDP—into rigorous documentation and structured datasets.