Academic CV · Updated 15 October 2026

Tanvir Ahmed Khan

Research Assistant, Machine Intelligence Lab
North South University · Dhaka, Bangladesh

Download CV (PDF, 2 pages) · Updated 15 October 2026

Education and research interests

B.Sc. in Computer Science and Engineering

North South University · 2021–2025 · Requirements completed December 2025

CGPA 3.80/4.00 · Summa Cum Laude

Relevant coursework: Machine Learning; Pattern Recognition; Numerical Methods; Probability and Statistics; Introduction to Linear Algebra; Design and Analysis of Algorithms.

Research interests

Multimodal learning; retrieval, memory, and evidence-aware reasoning; video understanding; 3D vision and neural rendering; audio-driven facial animation; low-resource natural language processing.

Ongoing research

ERL — Emergent Retrieval Landmark

I study how candidate access, evidence selection, and reading affect multi-hop answers. Preliminary findings motivate further comparisons with stronger retrieval and scoring methods.

Research ongoing · Manuscript in preparation · 2026–present

RIFT — Recoverability-Gated Implicit Facial Dynamics

I investigate how speech-predictable facial motion relates to rendered articulation. Following a negative predictive pilot, RIFT-M focuses on measurement reliability.

Active exploratory research · RIFT-M measurement study · September 2026–present

RSI — Capacity Growth in Verified Self-Improvement

I am building a verified self-training pilot to study model expansion. Reliable training targets and recovery checks come before testing the growth hypothesis.

Engineering pilot · Growth hypothesis untested · October 2026–present

Publications and manuscripts

Published and accepted

  1. 2026

    CAT-GS: Balanced Multimodal Learning via Calibrated Gating and Fusion Surgery

    Mahir Shahriar Tamim, Sharjil Khan, Md. Samiul Alim, Tanvir Ahmed Khan, Shafin Rahman, and Nabeel Mohammed.

    Neurocomputing, article 135392 (2026) · Published

    PublisherDOIarXiv

  2. 2026

    Can LLMs Follow the Pulse of a Crisis? Evaluating Crisis Sentiment in Bangladesh's July Uprising

    Md. Samiul Alim, Mahir Shahriar Tamim, Tanvir Ahmed Khan, Sharjil Khan, Rafia Ferdous Duti, Shahriyar Zaman Ridoy, and Mohammad Ali Moni.

    AACL-IJCNLP 2026, Main Conference · Accepted · Paper available

    Paper · OpenReviewarXivDataset page

  3. 2026

    Ensemble and temporal feature-based framework for rainfall classification in Bangladesh

    Mahir Shahriar Tamim, Md. Samiul Alim, Tanvir Ahmed Khan, Maisha Rahman, and Md Musfique Anwar.

    PLOS ONE 21(3): e0342646 · Published

    PublisherData

  4. 2025

    When a Nation Speaks: Machine Learning and NLP in People's Sentiment Analysis During Bangladesh's 2024 Mass Uprising

    Md. Samiul Alim, Mahir Shahriar Tamim, Maisha Rahman, Tanvir Ahmed Khan, and Md Mushfique Anwar.

    2025 28th International Conference on Computer and Information Technology (ICCIT), IEEE, pp. 413–418 · Published

    DOIarXiv

Research and industry experience

Research Assistant · Machine Intelligence Lab

North South University · Mar 2025–present

Advisors: Dr. Nabeel Mohammed and Dr. Shafin Rahman.

  • Contribute to multimodal learning, low-resource NLP, and speech-driven visual systems, including the publications listed above.
  • Developed distributed multi-GPU inference and chunked execution for collaborative talking-avatar systems; investigated Audio2Face–Three.js integration and ARKit blend-shape recovery.
  • Collaborated with Niloy on the PersonaPlex–IMTalker pipeline deployed and evaluated on Jetson AGX Orin, including streaming, renderer studies, and acceleration.
  • Profiling distinguished renderer acceleration from end-to-end latency. A research-manuscript direction is under exploration; client-side A/V timing and long-session behavior need further evaluation.

Junior Software Engineer · Tuctuc

Spain, remote · Jan 2022–Jun 2026

Maintained and extended Odoo-based ERP systems and developed cross-platform mobile applications using React Native.

Selected research and engineering projects

PNEUMA — Video Memory for Long-Form Question Answering

Completed December 2025 · Four-person team · Supervisor: Dr. Nabeel Mohammed

Integrated pretrained perception with SQL, multi-aspect vector retrieval, and temporal memory. Retrieval consistency and visual grounding require further evaluation. Research overview

Saliency Detection via Knowledge Distillation

Aug–Dec 2024 · Engineering prototype

Peekaboo/DINO teacher and UKAN student with mask-supervised and teacher–student objectives. GitHub

DynApex — Dynamic CPU/GPU Parallelization

Aug–Dec 2024 · Engineering prototype

C++ prototype for dependency and cycle analysis, pragma injection, and OpenMP-based CPU and target execution. GitHub

Skills, languages, and honors

Machine learning
PyTorch, Hugging Face Transformers, knowledge distillation, multimodal fusion.
Vision and avatars
3D Gaussian Splatting, FLAME/3DMM, NVIDIA Audio2Face, ARKit blend shapes.
Retrieval and memory
Retrieval-augmented generation, FAISS, SQL, temporal graphs, sentiment analysis.
Systems
Multi-GPU inference, TensorRT, Docker, Linux, NVIDIA Jetson AGX Orin, FastAPI, Git.
Programming
Python, C++, JavaScript, React Native.
Languages
Bangla (native), English (C1).
Honors and activities
Academic Scholarship, North South University (2023); participant, ICPC Asia Dhaka Preliminary Contest (2021).