Bikram Khanal
Applied Scientist advancing quantum machine learning and culturally aware language models.
About
I am an Applied Scientist at Amazon with a Ph.D. in Computer Science, specializing in quantum machine learning and large‑language models. I develop noise‑aware generalization theory and efficient optimization techniques for quantum machine learning on near‑term quantum hardware. I also design transparent and resource-efficient methods for evaluating language models, with a focus on mitigating locale biases in multilingual and culturally aware language models, particularly in conversational systems. I enjoy tackling complex problems and working across diverse machine‑learning domains.
Experience
Amazon.com Inc.
Applied Scientist II
Automated LLM-as-a-Judge system for assessing conversational quality of AI responses.
- Developed a scalable prompt-engineering and scoring framework that reduces manual review time for multilingual conversations.
- Designed quality assessment criteria covering multiple conversational dimensions to track model performance.
- Built an end-to-end evaluation pipeline leveraging LLMs as automated judges to inform iterative model improvements.
Skills: Prompt Engineering, Model Evaluation & Benchmarking, LLMs, Quality Metrics Design, Cross-cultural Localization
Amazon.com Inc.
Applied Scientist Intern
Advanced cultural awareness analysis for large language models.
- Evaluated prompt-engineering techniques that improve culturally localized responses.
- Researched verbosity bias in LLMs and quantified its impact on conversational performance.
- Optimized prompts to mitigate verbosity bias while preserving response quality.
Skills: Prompt Engineering, LLMs, NLP
Baylor University
Research Assistant
Established Tissue Bank for Alcohol Research
- Built and maintained a Django-based tissue bank application that streamlined bioinformatics data management and integration.
- Structured raw sample data into PostgreSQL to ensure integrity and improve researcher access.
- Created custom analytical functions enabling nuanced, time-sensitive investigations.
Skills: Python, PostgreSQL, Django
Improved Liver Disease Diagnosis
- Engineered a resource-efficient pipeline for 50–80 GB DZI image analysis to scale diagnostics.
- Visualized drinking patterns via Sankey diagrams, connecting gender, categories, and hourly ethanol consumption.
- Developed a sequential CNN in Keras achieving 98% accuracy for liver disease classification.
Skills: Image Processing, Sankey Diagrams, CNN, Optimization
Web Development
- Led Django platform enhancements that expanded functionality and user experience.
- Implemented security checks that hardened the platform against vulnerabilities.
- Containerized operations with Docker for reliable deployment and scalability.
Skills: Docker, Django, Python, Git, Remote Server Management
Baylor University
Teaching Assistant
Supported instruction across Computer Science curriculum.
- Led lab tutorials and assessments for C++, Python, Computer Architecture, Computing Fundamentals, and Algorithms.
- Partnered with faculty to design assignments and exams aligned with course outcomes.
- Mentored 150+ students through individual and group sessions to address learning challenges.
Skills: Interactive Learning, C++, Python, Curriculum Design, Teaching
Education
Baylor University
Baylor University
Troy University
Skills
Selected Publications (complete list)
- Exploring muzzle biometrics: a deep learning framework for noninvasive cattle recognition — Imaging Science: Computer Vision, Image and Signal Processing, Pattern Recognition (2025)
- Learning Robust Observable to Address Noise in Quantum Machine Learning — Communications in Computer and Information Science (2025)
- Data-dependent generalization bounds for parameterized quantum models under noise — The Journal of Supercomputing (2025)
- NLP-Guided Synthesis: Transitioning from Sequential Programs to Distributed Programs — Communications in Computer and Information Science (2025)
- Chapter 11 Exploring muzzle biometrics: a deep learning framework for noninvasive cattle recognition — Imaging Science (2025)
- A Modified Depolarization Approach for Efficient Quantum Machine Learning — Mathematics (2024)
- Automated Synthesis of Distributed Code from Sequential Snippets Using Deep Learning — 2024 IEEE International Conference on Big Data (BigData) (2024)
- Quantum-Enhanced Representation Learning: A Quanvolutional Autoencoder Approach against DDoS Threats — Machine Learning and Knowledge Extraction (2024)
- Generalization error bound for quantum machine learning in NISQ era—a survey — Quantum Machine Intelligence (2024)
- Evaluating the Impact of Noise on Variational Quantum Circuits in NISQ Era Devices — 2023 Congress in Computer Science, Computer Engineering, & Applied Computing (CSCE) (2023)
- Supercomputing leverages quantum machine learning and Grover’s algorithm — The Journal of Supercomputing (2023)
- Non-Invasive Muzzle Matching for Cattle Identification Using Deep Learning — 2023 Congress in Computer Science, Computer Engineering, & Applied Computing (CSCE) (2023)
- Attribution Scores of BERT-Based SQL-Query Automatic Grading for Explainability — 2023 International Conference on Computational Science and Computational Intelligence (CSCI) (2023)
- On Adversarial Examples for Text Classification By Perturbing Latent Representations — LatinX in AI at Neural Information Processing Systems Conference 2022 (2022)
- Human Activity Classification Using Basic Machine Learning Models — 2021 International Conference on Computational Science and Computational Intelligence (CSCI) (2021)
- Quantum Machine Learning: A Case Study of Grover’s Algorithm — 2021 International Conference on Computational Science and Computational Intelligence (CSCI) (2021)
Projects (Selected)
Locale- and Culturally-aware Response Generation in LLMs
Designed an LLM-as-a-Judge framework with context engineering and conversational metrics to evaluate and improve localized responses across cultures.
Generalization Error Bound for Quantum Machine Learning in NISQ Era--A Survey
A Systematic Mapping Study (SMS) to explore the state-of-the-art generalization bound for supervised QML in NISQ-era and analyze the latest practices in the field
A Modified Depolarization Approach for Efficient Quantum Machine Learning
A modified representation for a single-qubit depolarization channel using two Kraus operators based only on X and Z Pauli matrices.
Noise Evaluation on Variational Circuits
A thorough investigation of noise impact on quantum variational classification in the NISQ context over diverse dataset.
Muzzle Matching for Cattle Identification
A non-invasive muzzle matching to address the challenges in insurance fraud and animal trading markets.
Adversarial example generation using white-box attach on text embedding
A white-box adversarial attack on text embedding vectors through encoder-decoder model to generate adversarial examples.