Looking for a job

Seemal Tausif

@seemaltausif

Message

I build AI-driven security analysis and evaluation systems for complex software vulnerabilities.

Canada

Message

What I'm looking for

I'm looking to build AI-driven security, software analysis, or LLM evaluation systems where I can investigate complex vulnerabilities, automate reliable benchmarks, and translate research into practical impact.

At the University of Waterloo's CrySP Lab, I build LLM-augmented static analysis systems for Android security. My work improved prediction accuracy from 38% to 79%, analyzed 150+ services across 10+ OEM firmware images, and uncovered 32 access-control vulnerabilities in commercial Android ROMs.

At the National Research Council Canada, I engineered a Python module and automated evaluation workflow for agentic LLM systems. I curated multi-stage exploitation benchmarks, increased task success rates from 0% to 90%, and orchestrated 250+ sandboxed Docker evaluation pipelines.

I've also built static-analysis benchmarks at Waterloo's SWAG Lab and co-authored an INTERACT ’23 paper on an AI-powered postpartum-depression screening tool. I bring hands-on experience across Python, Java, LLMs, RAG, security research, and full-stack development.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer — Research Assistant

Current

University of Waterloo

Jan 2025 - Present (1 year 8 months)

Built an LLM-augmented static analysis pipeline to detect inconsistent security and device-configuration checks across Android system services, improving prediction accuracy from 38% to 79%. Discovered 32 access control vulnerabilities in commercial Android ROMs and received the Outstanding Poster Presentation Award at CPI Conference 2026.

Java WALA LLM RAG CodeT5+Android Security Static Analysis

Student Research Intern

National Research Council Canada

Jan 2026 - Aug 2026 (7 months)

Engineered a pluggable Python module within an agent-enhancement SDK, designing a test automation workflow using the Inspect AI framework to benchmark agentic LLM systems. Developed an integration benchmark of 5 multi-stage exploitation tasks, optimizing agent memory configurations to drive success rates from 0% to 90%.

Python Inspect Docker OpenRouter LLM Agentic Systems Benchmarking Testing Automation

Software Engineer — Research Assistant

Lahore University of Management Sciences

Sep 2022 - May 2024 (1 year 8 months)

Co-authored a peer-reviewed paper at INTERACT '23 presenting an AI-powered conversational screening tool for postpartum depression. Audited security and privacy compliance across 50+ children's gaming apps, identifying 40 COPPA violations and 25 deceptive design patterns.

AI Privacy Security Audit COPPA Research Python

Education

Degrees, certifications, and relevant coursework

University of Waterloo

Master of Mathematics, Computer Science

2025 -

Grade: 93/100

Activities and societies: International Master’s Award of Excellence (IMAE)

Pursuing a Master of Mathematics in Computer Science with a CGPA of 93/100. Coursework includes Systems & Software Security, Software Testing & Debugging, and Building Private Data Systems.

Lahore University of Management Sciences

Bachelor of Science, Computer Science

2020 - 2024

Grade: 3.94

Activities and societies: LUMS Merit Scholarship Award (2020-2024); Dean’s Honor List (2020-2024)

Completed a B.S. in Computer Science with a specialization CGPA of 3.94. Coursework included Linear Algebra, Probability, Discrete Math, Data Structures, Algorithms, Operating Systems, Machine Learning, Data Mining, Network Security, and Generative AI.