Choose a background

all three respond to your cursor

Divij Chawla

student / researcher / builder

I’m a University of Washington student studying computer science and physics, focused on AI security, model evaluation, and research systems.

Most of my recent work has been around automated red teaming, safety benchmarking, trace-based evaluation, and building technical infrastructure that holds up under real use.

I’ve worked across academic research, applied AI security, and systems-heavy engineering, and I’m especially interested in problems that are both technically deep and directionally important.

Focus

AI security ML: automated red teaming, safety benchmarking, monitoring, provider-SDK agent harnesses, trace-scored evals, reproducible PyTorch pipelines, Docker, and Slurm/HPC.

Education

University of Washington

Seattle, WA

Bachelor of Science in Computer Science · Expected 2028

GPA 3.80 · Dean’s List · Pursuing Double Major in Physics

Experience

Feb 2026 Present

SPAR Research Fellow (Shutdown-Bench)

Remote

Supervised Program for AI Research

  • Building a safety benchmark for LLM/agent shutdownability in realistic tool-use tasks: scenario suite, instruction hierarchy, and failure-mode taxonomy.
  • Implementing automated red-team agent harnesses with provider SDKs plus trace scoring to detect shutdown resistance, including delay, deflect, evasion, and goal-preservation behavior.
March 2025 August 2025

AI Security Researcher

Singapore, Remote

Walled AI

  • First author, EMNLP Industry Track 2025. Built FINRISKEVAL, a dataset and eval pipeline spanning 1,720 profiles, 8 models, and 13k+ outputs to measure correctness and intent alignment in finance, and ran 100+ prompt and scoring ablations.
  • Co-authored IMDA-commissioned LLM deployment playbooks covering red teaming, domain safety evals, intent alignment, and guardrails for production systems.
July 2024 March 2025

AI Research Intern

Bristol, UK, Remote

University of Bristol (UoB)

  • Built eval harnesses for latent alignment failures via representation backdoor attacks, fine-tuned attacker and defender models, tested mitigations, and quantified reliability under white-box and black-box conditions.
  • Threat-modeled 5+ attack vectors and measured sleeper-agent persistence and mitigation breakpoints across settings.
June 2023 July 2024

Software Development Intern

Bangaluru, India, Remote

Emsec Private Limited

  • Built and operated a honeypot fleet simulating 7k+ vulnerable applications to capture real-world attacker reconnaissance.
  • Investigated live attacker activity across several honeypots, curating datasets and labeling adversarial activity to strengthen risk-aware defensive automation.
  • Implemented a telemetry-to-evidence pipeline to parse, enrich, and aggregate attacker logs into usable detection signals.
May 2025 August 2025

DSP (Digital Signal Processing) Intern

Bangaluru, India

Emsec Private Limited

  • Optimized C++ DSP modules for high-throughput SDR pipelines and integrated real-time ingestion and processing for low-latency inference, with profiling and latency tuning.

Activities

  • Lavin Entrepreneurship Program
  • Interactive Intelligence (I2), UW
  • Husky Satellite Lab

Skills

Python, C++, PyTorch, Transformers, SQL, Linux, Git, Docker, Slurm/HPC, OpenAI Agents, Google ADK, evaluation harnesses, experiment tracking.