Daud Ibrahim Hassan | Empirical AI Safety Researcher
BRAC University · BSc CSE · Expected May 2028
Research
Featured research outputs
False Containment
Apart Research · AI Incident Response Sprint · September 2026 · Completed
I built the evaluation pipeline, verifier runs and evidence-ablation conditions for testing AI incident-resolution claims.
False Containment Rate: unresolved incidents incorrectly classified as resolved under incomplete telemetry. Author list remains to be verified.
Evidence ablation · Ground-truth verification · Completed sprint
When the Record and the Report Diverge
Apart Research · Digital Minds Research Sprint · August 2026
I built and ran the evaluation pipeline, checking model self-reports against deterministic tool-execution logs.
Haiku exact access-report fidelity: 96.5% → 57.6%. Free-form notes and Qwen retained 100% fidelity in this evaluation.
216 runs · 12 matched tasks · Two models
Research programs
AI Safety Camp
Research Incubator Participant · Remote · 2026 - Present
Machine unlearning research: reproducible adversarial evaluation protocols, Python analysis pipelines and recovery testing.
I test whether information targeted for removal remains recoverable under adversarial probing.
Research participation · Machine unlearning
Digital Minds Research Sprint
Apart Research · August 2026
Co-authored When the Record and the Report Diverge with Deniz Chen and Soumya Parthasarathy.
I built the evaluation pipeline and executed all 216 runs. Deniz Chen supported results calibration and Soumya Parthasarathy supported documentation.
LLM evaluation · Research sprint
AI Incident Response Sprint
Apart Research · September 2026 · Completed
False Containment: an empirical study of what evidence is sufficient to verify autonomous AI incident resolution.
My work covered the end-to-end evaluation pipeline, verifier runs, evidence ablation and ground-truth comparison. The publication author list remains to be verified.
Research sprint · Evidence verification
The Unlearning Illusion
BRAC University · Undergraduate thesis in progress
The Unlearning Illusion: A Red-Teaming Framework for Probing the Reversibility of Machine Unlearning in Large Language Models.
I’m developing adversarial recovery tests to distinguish apparent forgetting from genuine removal. The question is whether information targeted for removal remains recoverable under red-team probing. I build Python experiment and analysis pipelines for systematic comparison.
Machine unlearning · Adversarial recovery · Thesis in progress
MarkLens
Independent research · Sole-authored preprint · 2026
I built a reproducible SEC EDGAR pipeline that extracted 251,350 position-level holdings and constructed 2,163 same-loan, same-date cross-manager facility-quarters.
Matching was validated at 97.6% precision and 93.5% recall. Cross-manager valuation disagreement was 1.41 points of par, with a 95% confidence interval of 1.13-1.74. The code, panel dataset, analysis and Observatory were released with a DOI.
251,350 holdings · Reproducible pipeline · Independent preprint
Read the preprint and research release ↗Monitor Calibration Transport
Independent AI safety research · 2026 - Present
Monitor Calibration Transport: Evaluating AI Safety Monitor Reliability Across Models and Distribution Shifts.
I investigate whether safety monitors calibrated on one model or task remain calibrated across models, task distributions and deployment-like shifts. The focus is false-negative risk, transportability and the reliability of monitoring signals.
AI control · False-negative risk · Calibration transport
Research engineering beyond the paper
DeepLearning.AI · Course Tester · August 2026 - Present
I test pre-release AI/ML exercises, hands-on labs and learning content. I reproduce failures, check edge cases and provide structured technical feedback before release.
My open-source work also includes four merged upstream fixes across Grafana and Crawl4AI: template-variable interpolation, a panel-editor crash, Markdown structure and a WSL browser hang.
Failure analysis · Four merged fixes · Research tooling
Engineering
AI engineering & team delivery
Building Nexasity AI
Founder and AI/ML Engineer · Dhaka · August 2025 - Present
Grew the work from a solo founder effort to a three-person engineering team. I owned architecture, model integration, deployment, monitoring and client-facing delivery.
Shipped 10+ AI systems supporting 10,000+ registered users across the product portfolio and handling 37,000+ API requests per month.
10+ AI systems · 10,000+ registered users · 37,000+ monthly requests
Software & community tooling
Gregorians of North America
Official Website · Web Developer · Remote · October 2024 - Present
Developed and maintained a WordPress, PHP and JavaScript website serving 1,000+ monthly visitors.
Reduced page-load time by 40%, maintained 99.9% uptime and delivered 15+ front-end and back-end improvements.
Community website · Web development
Open source
Four merged upstream fixes
Grafana · Crawl4AI
Two upstream fixes in Grafana and two in Crawl4AI.
Grafana: Pyroscope template-variable regex interpolation and the Prometheus Both query-type panel-editor crash. Crawl4AI: Markdown heading/table structure and a WSL browser hang caused by a system proxy conflict.
Open source · Merged upstream
Technical collaborations
EquiStamp
Technical Contributor
Contributor experience at EquiStamp.
Technical collaboration
StampyAI / Alignment Research Dataset
Open-Source Contributor
Contributor experience with StampyAI / Alignment Research Dataset.
Open source · AI alignment
Leadership
Founder & team leadership
Building Nexasity AI
Founder and AI/ML Engineer · Dhaka · August 2025 - Present
Grew the work from a solo founder effort to a three-person engineering team. I owned architecture, model integration, deployment, monitoring and client-facing delivery.
Shipped 10+ AI systems supporting 10,000+ registered users across the product portfolio and handling 37,000+ API requests per month.
3-person team · 10+ AI systems
Community leadership & impact
WSEH English Speaking Platform
Host and Moderator · August 2021 - August 2023
Led 500+ sessions, organized events, maintained the server and handled moderation.
The work covered free speaking practice, coaching, session delivery and community oversight.
Volunteer leadership · 5,000+ members
Programming Hero
Campus Ambassador · March 2025 - September 2025
Organized seven hackathons for 200+ students and mentored 100+ students in machine learning and data science.
The profile reports a 40% increase in participation.
Seven hackathons · Campus community
BRACU Duburi AUV
Senior Member, Operations Sub-Team · 2025 - Present
Support operations and coordination for Bangladesh’s first autonomous underwater vehicle initiative.
I contribute to a team-built YOLO-based underwater navigation system, working on dataset collection, detector training, perception, tracking and vision-guided alignment with ROS2 and NVIDIA Jetson Orin Nano.
Team contribution · Operations · Robotics
Mentorship
Machine learning and data science mentorship
Programming Hero · Campus Ambassador · March 2025 - September 2025
Mentored 100+ students in machine learning and data science.
I also organized seven hackathons for 200+ students. The profile reports a 40% increase in participation.
100+ students mentored · Student community
Speaking practice and coaching
WSEH English Speaking Platform · Host and Moderator · August 2021 - August 2023
Led 500+ sessions, organized events, maintained the server and handled moderation.
I led a volunteer network of 50+ hosts and served as one of five moderators responsible for community oversight.
Community coaching · 50+ hosts
Robotics, hackathons & community activities
Underwater vision and navigation
BRACU Duburi AUV · Senior Member, Operations Sub-Team · 2025 - Present
A team project combining underwater perception, tracking and vision-guided alignment.
I collect datasets, train YOLO detectors and work with ROS2 and NVIDIA Jetson Orin Nano while supporting operations and coordination.
Robotics · Computer vision · Ongoing
Hackathons and student learning
Programming Hero · Campus Ambassador · March 2025 - September 2025
Seven hackathons organized for 200+ students, alongside mentorship for 100+ students in machine learning and data science.
The work connects technical learning with hands-on student projects and community participation.
Hackathons · Mentorship
Volunteer speaking community
WSEH English Speaking Platform · August 2021 - August 2023
Led 500+ sessions, organized events, maintained the server and handled moderation.
One of five moderators for a 5,000+ member English-speaking community.
Volunteer work · Community sessions
Gregorians of North America
Official Website · Web Developer · Remote · October 2024 - Present
Developed and maintained a WordPress, PHP and JavaScript website serving 1,000+ monthly visitors.
Reduced page-load time by 40%, maintained 99.9% uptime and delivered 15+ front-end and back-end improvements.
Community website · Web development
School & college clubs
Gregorian Science Club
Vice President / Festival Organizer · 2019 - 2020
Organized the 10th SGHSC Science, ART & Cultural Festival and the 11th SGHSC Science, Art, Cultural & Chess Festival.
Notre Dame Nature Study Club
Organizer / Executive Committee
Organized six NDNSC events between December 2020 and July 2022.
About
Research skills
Evaluation & AI safety
LLM Evaluation · AI Control and Monitoring · Red Teaming · Machine Unlearning · Deterministic Verification · Evaluation Harnesses · Benchmarking
Research methods & statistics
Reproducibility · Provenance and Leakage Controls · Failure Analysis · Probability Calibration · Confidence Intervals · Permutation Tests · Precision and Recall Validation
AI / ML skills
LLMs & generative AI
RAG · ChromaDB · Embeddings and Semantic Chunking · Grounded Generation · Prompt Engineering · LLM-as-Judge · Multi-LLM Routing · Multi-Agent Systems · CrewAI · LangChain · MCP
Machine learning & data
Python · PyTorch · scikit-learn · XGBoost · LightGBM · TensorFlow / Keras · FinBERT · Pandas · NumPy · Feature Engineering · Hyperparameter Tuning · Cross-Validation · SMOTE
On-device AI & robotics
Whisper · ONNX · Transformers.js · YOLO · ROS2 · NVIDIA Jetson Orin Nano · Detector Training · Perception and Tracking
Engineering skills
Backend, databases & security
FastAPI · REST APIs · SQL · PostgreSQL · Supabase · SQLite · JWT · OAuth 2.0 · Rate Limiting
Frontend & languages
React · Next.js · TypeScript · JavaScript · Tailwind CSS · Kotlin · PHP
Delivery & MLOps
Docker · GitHub Actions · CI/CD · Model Monitoring and Retraining · Azure · Vercel · Railway · Fly.io · Research Engineering
Analytics & visualization
Data Cleaning · Data Analysis · Web Scraping · Microsoft Excel · Tableau · MySQL · Microsoft Power BI
Background & education
BRAC University
BSc in Computer Science and Engineering · Expected May 2028
Dhaka, Bangladesh. My work spans AI/ML engineering, reproducible evaluation and technical AI safety.
My undergraduate thesis, The Unlearning Illusion, is in progress. It studies whether information targeted by machine unlearning remains recoverable under adversarial probing.
Education · Thesis in progress
Research engineering beyond the paper
DeepLearning.AI · Course Tester · August 2026 - Present
I test pre-release AI/ML exercises, hands-on labs and learning content. I reproduce failures, check edge cases and provide structured technical feedback before release.
My open-source work also includes four merged upstream fixes across Grafana and Crawl4AI: template-variable interpolation, a panel-editor crash, Markdown structure and a WSL browser hang.
Failure analysis · Four merged fixes · Research tooling
Certifications & courses
AI-Powered Software and System Design
DeepLearning.AI · Issued August 2026
Credential ID: FY60R7QFUGBE
AI System Design · AI Agents · Generative AI · Application Programming Interfaces (API) · Solution Architecture · Software Architecture · Systems Design · Prompt Engineering
Show credentialIntroduction to Social Media Marketing
Meta · Issued July 2026
Credential ID: 72I16TLOB111
Product Management · Performance Marketing · Marketing Management · Social Media Marketing
Show credentialGoogle AI Essentials Specialization
Google · Issued July 2026
Credential ID: E8HZC9JAZ3AR
Responsible AI · Artificial Intelligence (AI) · Generative AI · Prompt Engineering · AI Ethics
Show credentialCertificate of Completion: Model Context Protocol
Anthropic · Issued July 2026
Credential ID: ftcyjhjt9v8w
Model Context Protocol (MCP) · AI Agents · Python (Programming Language) · Generative AI · Large Language Models (LLM)
Show credentialCertificate of Completion: AI Fluency for Builders
Anthropic · Issued July 2026
Credential ID: bse9vehxnakg
Responsible AI · Artificial Intelligence (AI) · Generative AI
Show credentialData Analytics Bootcamp with Alex Freberg
Analyst Builder · Issued August 2025
Data cleaning and analysis using SQL, Excel and Python; visualization with Tableau and Power BI.
Data Visualization · Data Cleaning · Web Scraping · Microsoft Excel · Data Analysis · Tableau · MySQL · Microsoft Power BI
Journey
Monitor Calibration Transport
November 2026 · Planned · Research · Prepared; not formally started
Evaluating AI safety monitor reliability across models and distribution shifts.
StampyAI / Alignment Research Dataset
October 2026 · Open-Source Contributor
Contributor experience with StampyAI / Alignment Research Dataset.
EquiStamp
September 2026 · Technical Contributor
Contributor experience at EquiStamp.
False Containment
September 2026 · Research / AI Incident Response Sprint
Designed and implemented the end-to-end evaluation pipeline.
AI Safety Camp
August 2026 · Research Incubator Participant
Machine unlearning research: reproducible adversarial evaluation protocols, Python analysis pipelines and recovery testing.
When the Record and the Report Diverge
August 2026 · Research / Digital Minds Research Sprint
Built the evaluation pipeline and executed all 216 runs.
DeepLearning.AI
August 2026 - Present · Course Tester
Test pre-release AI/ML code exercises, hands-on labs, and learning content.
MarkLens
July 2026 · Independent research / preprint
I built a reproducible SEC EDGAR pipeline that extracted 251,350 position-level holdings and constructed 2,163 same-loan, same-date cross-manager facility-quarters.
The Unlearning Illusion
June 2026 · Undergraduate thesis in progress
Develop adversarial recovery tests that distinguish apparent forgetting from genuine removal.
Nexasity AI
August 2025 - Present · Founder and AI/ML Engineer
Designed and shipped 10+ AI systems spanning RAG, decision support, NLP, on-device inference, fintech, edtech, and developer tooling.
Programming Hero
March 2025 - September 2025 · Campus Ambassador
Organized seven hackathons for 200+ students.
Gregorians of North America
October 2024 - Present · Web Developer
Developed and maintained a WordPress, PHP, and JavaScript website serving 1,000+ monthly visitors.
BRACU Duburi AUV
January 2024 - Present · Senior Member, Operations Sub-Team
A team project combining underwater perception, tracking and vision-guided alignment.
WSEH English Speaking Platform
August 2021 - August 2023 · Host and Moderator
Led 500+ sessions, organized events, maintained the server and handled moderation.
Notre Dame Nature Study Club
18 December 2020 - 1 July 2022 · Organizer / Executive Committee
Organized six NDNSC events.
1 July 2022 · 13th National Nature Summit 2022
29 October 2021 · 12th National Nature Summit 2021-Hybrid Edition
29 October 2021 · Frame of Dreams: Season 4
5 June 2021 · 2nd Virtual Green Festival 2021: International Edition
22 April 2021 · Bangladesh Earth Summit 2021
18 December 2020 · 11th National Nature Summit 2020- Hybrid Edition
Gregorian Science Club
2019 - 2020 · Vice President / Festival Organizer
Organized the 10th SGHSC Science, ART & Cultural Festival and the 11th SGHSC Science, Art, Cultural & Chess Festival.
Contact
Interested in research or collaboration?
Email GitHub LinkedIn