.png){kind=logo}
Jimmy Dani
Mid-level AI Researcher specializing in privacy-preserving ML and applied cryptography
Connect with Jimmy
Jimmy already has a relationship with Reval, so a warm intro from us gets a much better response than cold outreach.
Recommended
Already have an account?
About
Graduate researcher who builds production-grade AI systems spanning LLM security evaluation and on-device RAG. Created HoneyLearner, a self-learning attack framework using GPT-4-class models as structured black-box attackers against honeywords defenses, with rigorous metrics and reproducible orchestration (Airflow/Spark/Kafka/Docker). Also partnered with agriculture scientists at Texas A&M–Corpus Christi to deliver UAV + 3D point-cloud crop-stress maps that cut time-to-insight ~40% and enabled ~30% earlier interventions.
Experience
Education
Key Strengths
- Built and deployed an LLM-driven security evaluation framework (HoneyLearner) for honeywords defenses
- Robust prompt+schema design to make black-box LLM outputs reproducible (strict JSON, validation, retries, fixed sampling)
- Quantitative evaluation mindset with custom metrics layer (top-k success, flatness, structural stats like Levenshtein/entropy/JS-divergence)
- Orchestrated reproducible ML/crypto pipelines with Apache Airflow (parallel multi-stage runs vs ad-hoc scripts)
- Designed and evaluated on-device RAG with measurable retrieval/answer quality plus latency/memory tuning
- Effective collaboration with non-technical stakeholders; delivered UAV/3D stress-mapping workflow adopted by agriculture scientists (40% faster insights, ~30% earlier interventions)
Discover more candidates like Jimmy
Search across thousands of pre-screened, high-quality, high-intent candidates on Reval.
Search Talent