Skip to content
View keerthanap8898's full-sized avatar
🦀
👩🏻‍💻
🦀
👩🏻‍💻

Block or report keerthanap8898

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
keerthanap8898/README.md

Keerthana Purushotham

◯ ☽ Computer Scientist ◐ Software Developer ◑ Research Engineer ❨ ☼

✧ Keerthana works at the intersection of AI, Distributed Systems, & Correctness; exploring how large-scale intelligent systems can be made more reliable, interpretable, & aligned with design intent.

  • Her work integrates research-driven inquiry with production-grade engineering. She currently works at AWS in the Threat, Security & Vulnerability Management team for Amazon-Linux ( AL12, AL1, AL2, AL2023, etc.) in EC2's Kernels & Operating Systems org ( KaOS )

✧ Keerthana has developed deep expertise in threat modeling & remediation, i.e, detecting new bugs & patching them; across more than 1,500 CVEs for multiple Amazon Linux (AL) distributions.

  • These threat detections & patches regularly touched every single one of the millions of AWS instances deployed globally including EC2 servers, AWS hypervisors, etc., during AL's fortnightly security releases.
  • Also involved orchestrating automated tests spanning various linux VM instances offered by AWS averaging monthly bills in the range of $15k - $50k / month for packages whose vulnerability lifecycles she's managed end-to-end;
  • This non-exclusively includes packages like docker, kernel, openssl, nss, python, java, mozilla, etc., amongst those seen in the AL2023 release notes : docs.aws.amazon.com/linux/al2023/release-notes/all-packages-AL2023.9.html, & more).

✧ She is a full-stack SDE with expertise in cybersecurity, cloud, NLP, & statistics. At AWS, she has integrated AWS CDK, C, Rust, Python, JavaScript, node.js, most major AWS tools & services, APIs, containers & shells, load-balanced edge-APIs & Lambdas running high frequency, global, federated, throttled workflows orchestrating async requests to collect critical threat data as soon as they're released & accurately evaluating them without delays to streamline & reliably execute engineering workflows & builds predictive automation tools for CVE evaluation, designs scalable cloud infrastructure, & supports threat detection for Amazon Linux.

✧ Keerthana's niche in AI, NLP, & computational statistics enables her to apply rigorous statistical methods to security analysis, threat modeling, & security R&D. She has also contributed significantly to system design efforts, ensuring that critical security information is incorporated effectively into real-world defenses. With a strong computer science foundation from UCSD, she has built skills across NLP, recommender systems, cloud architectures, & has published research. She is seeking impactful roles where she can drive innovation at scale.


Key Links ( x2 )

⎯⎯✧ ˚.☆°。𓆉 ྀ○°𓆝˚○。𓆡☆。⋆.݁݁✧˚𓆞。𓇼𓆝˚𓆟。༚⋅ ✧⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
⎯⎯✧ 𓋼𖧧˚°⚘𓃦。𓃙˚𓃠○𓃥°𓃚'⚘.𓏲˚𓍊𓋼✧。༚⋅ ✧⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

Networking:

⎯⎯✧ 𓂇𖧧𖠰ᨒ↟𓃬﹏↟𓂃𓃮ᨒ˚𖠰࣪↟𓃮﹏𓃮‿་༘ ✧⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

Focus Areas & Technical Interests:

Non-exclusively,

  • 🌕 AI System Reliability: Correctness & robustness in AI & distributed systems,
  • 🌔 Distributed Computing: Scalable, fault-tolerant architecture design,
  • 🌓 Scalable Machine Learning Infrastructure: Systems reasoning, verification, & interpretability,
  • 🌒 Program Analysis & Dataflow Optimization: Research-informed engineering practice,
  • 🌑 Systems for Alignment & Verification: Architecting provably-correct intelligent systems through formal methods, test-driven reasoning, and algorithmic accountability.
⎯⎯✧ 𓇢𓆸𓇗⚘𖤣𖥧𓏲°✾.𓅰.𓅭.𓅮.𓅯.𖡼˚↟𖠰✧𖤣𖥧𓅪𖧧⋆ ✧⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

Career:

She is open to conversations & new opportunities around AI systems research, reliability engineering, & correctness-oriented design.

⎯⎯✧﹌𓆤༉𖧧𖥧𖤣. ༘༝ၴ( ၴႅၴ˖𓏲⚘ཐི༏ཋྀˎ ྀ𓏲𓇗𖤣﹏𓆏࿐⚘𖥧𖤣𓇗ˎˊˎˊ𓆈ˊˎ゛✧⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
Sl.# Category Links
❶. Matrix ( fedora ) / Pagure @keepur:fedora.im / fedoraproject.org/wiki/User:Keepur / pagure.io/user/keepur
❷. Fedora / Redhat accounts.fedoraproject.org/user/keepur / access.redhat.com/account/57599301
❸. Website ( personal ) / LinkedIn keerthanap8898.github.io/keerthanap8898 / linkedin.com/in/keerthanapurushotham
❹. GitHub / github-Bio / github-Repositories github.com/keerthanap8898 / github.com/keerthanap8898/bio / github.com/keerthanap8898?tab=repositories
❺. Mastodon / Bluesky @keepur@infosec.exchange / @keepur8.bsky.social
❻. Google-Scholar / ResearchGate scholar.google: user=tWzF13sAAAAJ / ResearchGate: Keerthana Purushotham
❼. Medium / Substack Medium: @keerthanapurushotham / Substack: @keerthanapurushotham
❽. X ( twitter ) / Discord X: keepur8 / Discord: 747152507184349195 - ( keepur8 )
❾. AI Chatbot notebooklm.google.com/notebook/fe2125af-e6e0-4815-8181-041b267e3b8b

P.S.

Someday she'll quit messing with unicode symbols. Not today though.


Pinned Loading

  1. bio bio Public

    This is where it's all at 💫 (˙ᵕ˙) ྀི —— Scroll down to the README

  2. CveToad CveToad Public

    WIP to help catch bugs and whatnot

    2

  3. TextToVideoAPI TextToVideoAPI Public template

    Async text2video API with the Genmo Mochi-1 model on 8×H100 GPU K8 worker nodes. Backend will handle job submission, tracking, & retrieval via JSON endpoints. Basic frontend for prompt submission, …

    Rust

  4. NeuralCRFs_for_ConstituencyParsing NeuralCRFs_for_ConstituencyParsing Public

    I present the CKY algorithm for constituency parsing on the PTB_LE10 & PTB_first_2000 subsets plus the full PTB dataset; to compute model features needed to calculate the partition function in the …

    Jupyter Notebook

  5. Neural-CRF_NER-Tagger Neural-CRF_NER-Tagger Public

    How to build a baby-BERT : I analyze BiLSTMs combined with Conditional Random Fields for Named Entity Recognition & contrasts a Neural-CRF tagger against a baseline BiLSTM model, exploring how prob…

    Jupyter Notebook

  6. context-based-comment-filtering context-based-comment-filtering Public

    This repository contains a sub-project to filter out irrelevant comments based on the context of the initial search in a data-set of web scraped comments and replies from social media websites.

    Jupyter Notebook