Tahmid Khan

Sep 2026 – Present

• Leading adversarial quality assurance (QA) loops to stress-test frontier models against complex multi-turn jailbreaks, prompt injections, and safety guardrail bypasses.
• Auditing expert-level adversarial data submissions to eliminate label noise, reconcile contradictory evaluation matrices, and enforce strict rubric compliance.
• Engineering high-complexity technical content and alignment benchmarks, validating structural reasoning and linguistic logic constraints across enterprise training datasets.

Jun 2026 – Present

• Adversarial red-teaming and model-breaking campaigns for flagship frontier models, uncovering critical failure modes in multimodal reasoning.
• Constructed complex prompt injection and test scenarios with deterministic ground truths to eliminate subjective bias and standardize vision-language performance metrics.
• Engineered complex image-text contradiction tests using strategically sourced CC0 assets to evaluate model robustness, safety guardrails, and factual grounding.

Jan 2026 – Present

• Evaluating autonomous agent execution loops, multi-step tool-use pipelines, and context-dependent routing logic to optimize multi-agent system reasoning.
• Stress-testing code-generation and text models against edge cases, tracking multi-turn hallucination decay, and verifying deterministic factual grounding.
• Reviewing and optimizing dense, rubric-driven evaluation datasets to align model outputs with safe, human-centered logic constraints.

Feb 2026 – Sep 2026

• Multimodal Evaluation: Conducted side-by-side comparative analysis of advanced vision, audio, and language models; systematically refined evaluation rubrics for image/video synthesis and audio quality metrics.
• Computer Vision & Segmentation: Orchestrated precise video segmentation, object tracking, pointing, and hard-reference validation to enhance model comprehension of spatial-temporal boundaries.
• Media Manipulation & Editing: Curated targeted datasets for generative video tasks, validating model logic during complex object addition/removal pipelines and automated captioning loops.

Aug 2024 – Aug 2025

• Full-Cycle Content Development: Authored comprehensive technical reports and research modules focused on Al and Computing Sciences, managing the entire lifecycle from introduction to conclusion.
• Data Visualization: Designed and integrated complex data graphs and charts using MS Word to illustrate research findings and technical trends.
• Digital Research: Conducted deep-dive secondary research using global academic databases to synthesize information for diverse technical projects.

Mar 2023 – Jul 2024

• Evaluated and refined structural logic in technical reports, ensuring clarity, flow, and adherence to formal documentation standards.
• Facilitated remote workshops on advanced research techniques and information synthesis, focusing on accuracy and analytical depth.
• Optimized communication workflows by training others to transform raw data into professional executive presentations (PPTX).

Feb 2022 – Feb 2023

• Promoted to Quality Reviewer: Elevated to the 'Remo Reviewer' tier for maintaining superior accuracy metrics; audited historical submissions of entry-level annotators to eliminate label noise. Remo Reviewer ID: 61cb14350...
• Sensor Fusion & LiDAR: Annotated complex 3D spatial environments utilizing LiDAR cloud points and multi-camera data to train autonomous vehicle machine learning perception systems.
• Legacy Computer Vision: Designed manual 2D bounding boxes and tracking matrices for complex video sequences, establishing early baseline datasets for computer vision models.

M.Sc in Computing Sciences (Major in AI and Data Engineering)

University of Vaasa

Aug 2025 – Aug 2027

B.Sc in CSE (Computer Science and Engineering)

Daffodil International University

Jan 2018 – Aug 2022