summaryrefslogtreecommitdiff
path: root/cv/index.txt
diff options
context:
space:
mode:
authorCaptainJack2491 <jayrupnakawala@gmail.com>2026-09-02 11:00:42 +0100
committerCaptainJack2491 <jayrupnakawala@gmail.com>2026-09-02 11:00:42 +0100
commit0ee667ac6cb2493e4f9ddaa5b69f595071aca936 (patch)
tree9051ebe3279544efe4ba228aa311b6bf41d4f7d3 /cv/index.txt
parente7dc4befde5bef83d198a631926174bae612f866 (diff)
fix: title frontmatter, terminal resume, project list sync, nginx hardening, and stylesfixes/titles-nav-typos
Diffstat (limited to 'cv/index.txt')
-rw-r--r--cv/index.txt67
1 files changed, 66 insertions, 1 deletions
diff --git a/cv/index.txt b/cv/index.txt
index 5f4b42e..efd50a9 100644
--- a/cv/index.txt
+++ b/cv/index.txt
@@ -1 +1,66 @@
-use `curl jayrup.me/cv.pdf > jayrup_cv.pdf`
+JAYRUP NAKAWALA
+London, UK | jayrupnakawala@gmail.com
+jayrup.me | github.com/CaptainJack2491 | linkedin.com/in/jayrupnakawala
+------------------------------------------------------------------------
+
+RESEARCH EXPERIENCE
+------------------------------------------------------------------------
+
+[1] Prime-Grokking: Pre-registered Grokking Study (Aug 2026 — Present)
+ git.jayrup.me/c/prime-grokking.git
+ - Tested whether minimal transformers and RNNs can grok next-prime prediction,
+ running the study like production research: every experiment batch pre-registered
+ with commit-locked interpretations before launch.
+ - Caught two experimental-design blockers via hostile pre-run review; built an
+ idempotent multi-cell sweep runner and batched the evaluation loop 46x faster.
+ - Ran a 16-cell PyTorch CUDA/AMP sweep (architecture x weight decay) on a remote GPU
+ node, then extended through 10x more data and a 20x longer horizon.
+ - Published a 3-part writeup at jayrup.me with a public, re-runnable repository.
+
+[2] Stress-Testing the J-space Interpretability Lens (Jul 2026 — Aug 2026)
+ jayrup.me/blog/jacobian-lens-frequency
+ - Independently reproduced a frontier interpretability method (Anthropic's July 2026
+ J-space/J-lens analysis) on a small PyTorch transformer and found scores track token frequency.
+ - Designed a matched-frequency control showing contextual predictability remains
+ detectable — framed and published as a control requirement for the method.
+
+[3] Dissertation: Moral Cover and Watchful Eyes (Nov 2025 — May 2026)
+ github.com/CaptainJack2491/dissertation | jayrup.me/dissertation
+ - Architected a sandboxed evaluation framework with a Virtual File System (VFS) to safely
+ execute and monitor autonomous LLM tool-use across multiple providers.
+ - Built a provider abstraction layer for glass-box reasoning extraction and post-hoc
+ interrogation to probe agent intent.
+ - Designed a 3-prong automated judging pipeline and ran 960 experimental trials.
+ - Quantified a "Stealth Gap": deceptive intent held steady (~62–67%) under high oversight
+ while output-level detectability dropped (47.8% -> 35.6%).
+
+PROJECTS
+------------------------------------------------------------------------
+
+[1] Hybrid-Cloud Research Infrastructure & Zero-Trust Mesh (Jan 2024 — Present)
+ - Operate a multi-node Linux estate (Void, Arch, Ubuntu) unified over a Tailscale/WireGuard
+ zero-trust mesh.
+ - Run 15+ containerized services via Docker Compose with automated backups, resource
+ monitoring, and an NVMe-backed storage tier.
+ - Solved CGNAT reachability using VPS forward proxies without port forwarding.
+ - Built jayrup.me on this stack: Nginx dual-interface serving (HTML vs plaintext) with a Quarto
+ pipeline rendering web and CLI docs from one Markdown source.
+
+EDUCATION
+------------------------------------------------------------------------
+
+University of East London (London, UK)
+BSc (Hons) Data Science & Artificial Intelligence — First-Class Honours (94%) (Sept 2023 — May 2026)
+Relevant Coursework: Math for Computing, Data Structures & Algorithms, Networking, Databases, Big Data, AI
+
+SKILLS
+------------------------------------------------------------------------
+
+Programming: Python, Bash, SQL
+ML / AI: PyTorch, Transformers, Hugging Face, llama.cpp (GGUF quant), LangChain, scikit-learn
+Infrastructure: Docker, Docker Compose, Nginx, Caddy, Tailscale, WireGuard, Quarto
+Systems: Linux (Arch, Void, Ubuntu), Git, rsync, Tmux, Neovim
+
+------------------------------------------------------------------------
+PDF DOWNLOAD: curl -O https://jayrup.me/cv.pdf
+------------------------------------------------------------------------