Latest EMNLP 2026 Main
DSAgentBench
Can agents automate end-to-end data-science workflows in real computer environments?
275 tasks 15 models deterministic evaluators 56.70% best result
Applied machine learning
Building reliable AI agents, multimodal systems, and model evaluation infrastructure.
At the Vector Institute, I turn applied ML research into production-ready systems, from benchmark design and model training to deployment and evaluation.
- Efficient Inference
- LLM Evaluation
- AI Agents
- Post-Training
- Multimodal AI
Selected work
Selected systems and benchmarks
Projects across agent evaluation, multimodal learning, post-training, and visualization.
DSAgentBench
A benchmark for evaluating whether AI agents can complete end-to-end data-science workflows in real computer environments using notebooks, IDEs, terminals, browsers, and databases.
Evidence: 275 diverse tasks across the data-science lifecycle, 15 evaluated models, and deterministic evaluators. The best reported agent reached 56.70% task success.
Fine-Grained Benchmark Generation
A reference-grounded multi-agent pipeline for generating, verifying, repairing, and deduplicating technically demanding evaluation tasks.
Contribution: Built OCR ingestion and a multi-agent designer and verifier workflow with iterative repair, trace-aware validation, and semantic deduplication.
Evidence: Trace-aware quality control connects source ingestion to final benchmark validation.
Open-PMC-18M
An open multimodal dataset and scalable pipeline linking biomedical figures with captions and subcaptions for medical vision-language learning.
Contribution: Owned preprocessing and batched vLLM inference for Qwen2.5-VL and Qwen2.5 summary generation.
Evidence: Scales compound-figure extraction, summarization, and OpenCLIP training across 18 million biomedical image-text pairs with optimized batched VLM inference.
RL-Text2Vis
A GRPO framework that evaluates text-to-visualization outputs after code execution using textual, executable, and visual feedback.
Contribution: Co-developed post-execution multimodal feedback for aligning generated text, code, and rendered visualizations.
Evidence: Improves code executability and chart quality over strong prompting and supervised baselines.
DashboardQA
A benchmark for agents that answer questions through real interactive dashboards.
Contribution: Co-created the benchmark to expose failures in grounding, planning, GUI interaction, and multi-view reasoning.
Evidence: Tests filters, tabs, view switching, visual grounding, and cross-view reasoning in realistic dashboard environments.
ChartQAPro
A challenging chart question-answering benchmark built around diverse real-world visualizations.
Contribution: Co-created the benchmark and evaluation protocol for open and closed vision-language models.
Evidence: 1,341 charts and 1,948 questions evaluated 21 VLMs.
Research
Selected publications
Recent work on AI agents, multimodal evaluation, data storytelling, and reliable generation.
-
EMNLP Accepted
DSAgentBench: Can Agents Automate End-to-End Data-Science Workflows in Real Computer Environments?
EMNLP 2026 Main
-
EACL Conference
Aligning Text, Code, and Vision: A Multi-Objective Reinforcement Learning Framework for Text-to-Visualization
-
EACL Findings
DashboardQA: Benchmarking Multimodal Agents for Question Answering on Interactive Dashboards
-
ACL Findings
ChartQAPro: A More Diverse and Challenging Benchmark for Chart Question Answering
-
EMNLP Conference
DataNarrative: Automated Data-Driven Storytelling with Visualizations and Texts
-
EMNLP Industry
Deploying Tiny LVLM Judges for Real-World Evaluation of Chart Models: Lessons Learned and Best Practices
Career
Experience and education
Roles in applied ML research, engineering, and teaching.
2025-Present
Vector Institute
Associate Applied Machine Learning Specialist · Toronto
2023-2025
Intelligent Visualization Lab, York University
Graduate Research Assistant · Toronto
2021-2023
Islamic University of Technology
Lecturer · Bangladesh
MSc Computer Science - York UniversityBSc Computer Science and Engineering - Islamic University of Technology
Now
Recent milestones
Recent publication and career updates.
| Aug 2026 | OpenPMC-18M was accepted at MICCAI 2026. |
|---|---|
| Aug 2026 | DSAgentBench, a benchmark for end-to-end data-science workflows in real computer environments, was accepted to EMNLP 2026 Main. |
| May 2026 | Released the preprint Fine-Grained Benchmark Generation for Comprehensive Evaluation of Foundation Models, describing grounded task generation with multi-agent design and verification. |
Service
Leadership and reviewing
President of the Bangladeshi Graduate Student Association at York and reviewer for NeurIPS, ACL, EMNLP, COLM, NAACL, and ACL Rolling Review.