Rakib Kibria is a Full-Time Lecturer in the Department of Computer Science and Engineering at BRAC University. He received his B.Sc. in CSE from BUET in March 2025. His research lies at the intersection of computational geometry and data-driven methods, focusing on learned spatial query processing and the approximation of geometric predicates such as visibility. His broader interests include geometric reasoning under uncertainty, differentiable geometry, spatial optimization, and the evaluation of learned spatial systems.

Courses Taught:

• CSE 420 - Compiler Design
• CSE 421 - Computer Networks

B.Sc. in Computer Science & Engineering, Bangladesh University of Engineering & Technology (BUET)
Passing Year: 2025
CGPA: 3.95 / 4.00

Higher Secondary School Certificate (HSC), Notre Dame College, Dhaka
Passing Year: 2019
GPA: 5.00 / 5.00

Secondary School Certificate (SSC), Bogra Zilla School, Bogra
Passing Year: 2017
GPA: 5.00 / 5.00

University Merit Scholarship, Bangladesh University of Engineering and Technology (BUET)

Dean’s List Scholarship, Bangladesh University of Engineering and Technology (BUET)

Talent Pool Scholarship, SSC 2017 - 4'th place in the Rajshahi Board

1. Learned Spatial Query Processing

Research on neural approximation of geometric predicates in spatial databases, with emphasis on:

• Visibility and proximity queries under high obstacle density
• Learning-based spatial query processing and geometric reasoning
• Neural approximation of exact spatial operators

Information-Theoretic Analysis of Spatial Representations:

•  Quantifying information loss introduced by discretized and rasterized scene representations
•  Deriving upper bounds on achievable prediction accuracy from representation constraints
•  Studying the relationship between spatial resolution and query accuracy

2. Geometric Learning Under Uncertainty

Research on spatial predicate evaluation under uncertain and noisy geometry, particularly

Uncertainty-Aware Geometric Reasoning:

• Learning spatial predicates from noisy or partially observed obstacle geometry
• Studying cases where exact geometric algorithms produce confidently incorrect results due to erroneous inputs
• Propagating input-geometry uncertainty through downstream spatial analyses

Calibrated and Distribution-Free Prediction:

• Calibrated probabilistic outputs for geometric queries
• Application of conformal prediction to obtain distribution-free coverage guarantees
• Quantifying uncertainty in learned visibility, reachability, and spatial predicates

3. Differentiable Geometry & Spatial Optimization

Research on differentiable formulations of geometric problems, enabling gradient-based optimization for traditionally discrete or combinatorial tasks.

Differentiable Geometric Predicates:

• Development of smooth surrogates for discontinuous geometric predicates
• Enabling gradient-based optimization for visibility, intersection, and spatial constraints
• Studying the relationship between surrogate objectives and exact geometric formulations

Spatial Optimization:

• Facility placement and sensor coverage optimization
• Guard positioning and visibility-based optimization
• Differentiable formulations benchmarked against exact combinatorial approaches

Continuous Relaxation of Discrete Problems:

• Analysis of solution quality and convergence under continuous relaxation
• Trade-offs between computational efficiency and optimality guarantees
• Characterization of when differentiable optimization provides practical advantages over discrete search