Katharine Jarmul
where else to find them
x:kjam
linkedin:katharinejarmul
what they talk about
all 10 in the filters
where they keep going back
- GOTO Copenhagen2 editions2024–2025
- Strange Loop2 editions2018–2022
- QCon São Paulo2 editions2018–2019
all 13 in the filters
on a stage with
- Emily Gorcenskionce
- Marielle Dadoonce
- Matteo Guzzoonce
all 6 in the filters
who talks about the same things
- Jigyasa Grover8 talks
- Vincent D Warmerdam22 talks
- Tess Ferrandez-Norlander24 talks
- Joshua Arvin Lat20 talks
- Jodie Burchell18 talks
- Rishabh Misra5 talks
| title | conference | date |
|---|---|---|
| Hacking AI Systems: How to (Still) Trick Artificial Intelligence | GOTO Copenhagen 2025 | 2025-09-29 |
Survey of adversarial ML shows concrete exploits and defenses: training-set extraction reproduces images (example: Carlini reconstruction of training images… | ||
| Unforgettable, that's what you are: Evaluating Machine Unlearning and Forgetting | PyCon DE & PyData 2025 | 2025-04-23 |
Shows that large models memorize both common and novel (outlier) examples and that attackers can use shadow models + discriminators for membership inference… | ||
| Encrypted Computation: What if Decryption Wasn’t Needed? | GOTO Copenhagen 2024 | 2024-09-30 |
Live demo encrypts a sentence with Zama on a Hugging Face space, runs an XGBoost inference and returns a decryptable neutral sentiment result, showing… | ||
| DIY Personalization: How, when why to offer self-made AI | PyData Paris 2024 | 2024-09-25 |
A demo builds a local RAG in a Jupyter notebook: initial LLM-as-retrieval fails, switching to BM25 + sentence encoder plus reranking and a local LLaMA-style… | ||
| Practical Data Privacy | GOTO Amsterdam 2023 | 2023-06-13 |
A practical claim: pure anonymity is mathematically impossible, so systems must trade information for privacy and choose techniques deliberately. A Netherlands… | ||
| If only I owned my data: Architecting decentralized data | Strange Loop 2022 | 2022-09-22 |
Slides argue a concrete architecture: decouple identity, storage, access and ML so users keep keys and move data; cite working pieces like did standards and… | ||
| Some useful bits: A few solved development problems via Crypto | XConf Europe 2022 | 2022-07-20 |
Covers concrete fixes: use private set intersection (Diffie‑Hellman and newer PSI/OPRFs) to join IDs without sharing emails, use secure multi‑party computation… | ||
| Decentralized identity: What’s all the fuss about? | XConf Singapore 2022 | 2022-07-15 |
Defines how decentralized identity actually works: a DIDs-backed wallet holds keys and produces verifiable credential and verifiable presentation pairs that… | ||
| Career Panel | PyCon DE & PyData Berlin 2022 | 2022-04-11 |
Panel records concrete tradeoffs: PhD pays off only for a topic one loves and funded positions matter; freelancing means being hired as an instant expert on… | ||
| Keynote - Big Data's Ethical Drought | QCon São Paulo 2019 | 2019-05-06 |
Argument: the rush for data produces concrete harms and sloppy choices. Examples: viral WhatsApp images spread false election claims at scale; classic datasets… | ||
| Computers are Stupid: Protecting "AI" from Itself | GOTO Berlin 2018 | 2018-10-31 |
Adversarial examples repeatedly turn real objects into high‑confidence misclassifications (a 3D‑printed turtle gets labeled a rifle at 90% confidence)… | ||
| Privacy: the Last Stand for Fair Algorithms | Strange Loop 2018 | 2018-09-26 |
Argues that privacy protects fairness by design and shows concrete methods: differential privacy as a formal guarantee, learned fair representations that… | ||
| Opening Keynote | PyCon APAC 2018 | 2018-05-31 |
A live demo shows a Python bird-or-not classifier implemented in under 50 lines using a Keras-style library and ImageNet-pretrained weights, illustrating… | ||
| Ethical Machine Learning | QCon São Paulo 2018 | 2018-05-09 |
Covers concrete failures and fixes: predictive policing simulations show feedback loops that concentrate stops until under-policed areas vanish, and an… | ||
| Keynote: Ethical Machine Learning: Creating Fair Models in an Unjust World | PyData Amsterdam 2017 | 2017-04-07 |
Claims that real systems reproduce societal bias with concrete examples and fixes: ProPublica COMPAS finding shows risk scores correlate with race; Google… | ||
| Holy D@t*! How to Deal with Imperfect, Unclean Datasets | PyData London 2016 | 2016-05-06 |
Specific, actionable tactics appear: use pandas backfill/ffill/fill for time series gaps and watch for hidden nulls like repeated nines; detect duplicate… | ||
| nothing matches | ||