Have a Question about JIIT?
Chat with VIDYA
+91-9999912862+91-9999962128
searchImportant Announcements:
Admissions Open for 2026.Apply NowAnother Opportunity : Open House (Parent Interaction) on 13th June 2026.Register NowCareer OpeningsApplyRound-1 of 10+2 Marks Based Counselling Scheduled for 03 June 2026. Instructions
Academic Experts
Baibhav Singh

Biography

Baibhav Singh is a Faculty Fellow in the Department of Computer Science and Information Technology at the Jaypee Institute of Information Technology (JIIT). He earned his Bachelor of Technology (B.Tech.) in Computer Science in 2025 from JIIT’s Sector-62 campus in Noida. His research spans green computing, quantum computing, transformer-based NLP (e.g., ChatGPT), software code quality, machine learning, recommender systems in software engineering, computer vision, and explainable AI. His work on problem solving, data structures, and algorithms further underpins his commitment to advancing the state of the art in computing. 

An enthusiastic competitive programmer, Baibhav hones his algorithmic and problem-solving skills through platforms such as CodeChef and Codeforces. Passionate about sustainable and responsible technology, he endeavors to integrate eco-friendly practices into all aspects of his work. His multifaceted interests—combined with a dedication to excellence—position him as a dynamic educator and researcher who contributes significantly to both the academic community and the broader field of computer science.
 

Research Highlights

Baibhav Singh developed a transformer-based code analysis framework that prioritizes both efficiency and interpretability. The system leverages pretrained models (CodeBERT and GPT-2) to detect bugs and suggest refactorings in Python code. An attention-visualization layer highlights relevant tokens (e.g., variable names or loop constructs) for each diagnostic, providing transparent explanations of the model’s predictions. To handle large codebases, the tool splits code into chunks and employs sparse or quantized transformer variants, which significantly speed up processing without degrading accuracy. Experimental results demonstrate the approach’s practicality: the model reliably detects complex errors (such as undefined variables and logic flaws) and generates context-aware optimizations (e.g., loop simplifications and expression pruning). 

Singh’s work shows that adversarial training and efficient attention mechanisms can yield a responsive, accurate code analysis service. Future plans include expanding beyond Python, integrating the tool into IDEs for real-time feedback, and adding advanced code-smell detection to further improve developer productivity. 

In a second study, Singh performed an empirical evaluation of quantum algorithms using the Cirq simulator, implementing representative quantum routines (Shor’s factoring, Grover’s search, the HHL linear solver, QAOA for optimization, and the BB84 quantum key distribution protocol) alongside their classical counterparts. The experiments confirmed the algorithms’ theoretical advantages on small instances: for example, Shor’s circuit factored small integers (N = 15, 21), while Grover’s search exhibited the expected √N speedup over brute-force search. The HHL solver encoded solutions with over 99% fidelity for a 2×2 system, and the BB84 protocol securely generated encryption keys with interception detection. By contrast, the 4-node QAOA did not outperform classical heuristics at this scale. Overall, Singh’s findings reinforce that quantum methods hold transformative potential (e.g., exponential factoring and quadratic search gains) but that current hardware and simulators impose overheads; classical algorithms remain faster on small problems. Singh suggests that future work should focus on scaling these circuits, overcoming noise and resource limits, and refining algorithms so that full quantum advantages become practical as larger, fault-tolerant quantum computers emerge.
 

Areas of Interest
  • Efficient and Sustainable Computing
  • Quantum Computing
  • Cyber security
  • Explainable Artificial Intelligence (XAI)
  • Machine learning
  • Data Structures and Algorithms
Publications
  1. H. Mathur, B. Singh, R. Sandilya, and S. K. Singh, “Efficient and interpretable code analysis with transformers,” Proc. 2025 International Conference on Contemporary Computing (IC3), Noida, India, 2025, in press.
  2. H. Mathur, R. Sandilya, B. Singh, and S. K. Singh, “Empirical study of quantum algorithms,” Proc. 2025 International Conference on Contemporary Computing (IC3), Noida, India, 2025, in press.