Arjun Desikan Fig. 00 Composite List
Fig. 00 — Composite

All six inks

36 items · PC1 9% / PC2 8% of variance
Shape ID Title Blurb Layer Year PC1 PC2
R-01 AI Safety & Agent Harness Research Fellow
Theorem.dev · Apart Research
 LLM-driven specification synthesis that turns C and Rust code into machine-checked proofs in Frama-C and Lean 4. Research 2026 0.47 −0.86
R-02 Quantitative Finance Researcher
MIT CSAIL
 Representation learning for markets: a 4.9M-parameter time-series transformer that learns temporal and cross-asset structure across 499 stocks. Research 2026 0.40 0.63
R-03 Scientific Machine Learning Researcher
Johns Hopkins · Centrum IntelliPhysics
 Neural operators that predict touch from vision: DeepONet and Fourier DeepONet models for visual-to-tactile acceleration. Research 2026 0.45 0.62
R-04 Student Researcher (First-Author)
Harvard OpenBio
 First-author study linking copy-number alteration burden to transcriptomic complexity across 3,527 tumors in 12 cancer types. Research 2025 −0.12 0.27
E-01 Software Engineer & Computational Finance Researcher
MIT CSAIL
 Shipping features for Mantis, the lab's swarm-intelligence software, alongside research on financial representation learning. Experience 2026 0.38 −0.15
E-02 Data Science Intern
Fort Bend County Health Dept.
 Mortality statistics and respiratory-disease forecasting for a Texas county health department, in a brief the county used to inform policy. Experience 2025 0.13 0.56
P-01 Agent Ledger: Shared Memory for AI Agents A shared memory layer that lets web agents learn from past failures, and a benchmark showing when retrieval helps and when it hurts. Projects 2026 0.51 −0.55
P-02 GitXL: Version Control for Excel Git for Excel workbooks: cell-level diffs, three-way merges, and a normalizer so re-saving an unchanged file produces no diff. Projects 2026 −0.28 −0.05
P-03 Monte Carlo ICU Risk Model An absorbing Markov chain fit to 14,000+ ICU transitions, with a Monte Carlo engine for tail risk in mortality and length of stay. Projects 2026 0.33 0.27
P-04 Glitch Market An on-chain marketplace for speedrun glitches, with encrypted delivery, decaying prices and jury-settled disputes. Projects 2026 −0.29 −0.49
P-05 ECG Classification with a 1D CNN A 1D CNN that flags abnormal ECGs at 0.98 AUC and uses Grad-CAM to show which parts of the heartbeat drove each call. Projects 2026 0.32 0.65
P-06 Equities Backtesting Dashboard A Streamlit dashboard for backtesting EMA-crossover and mean-reversion strategies, stress-tested with Monte Carlo resampling. Projects 2026 0.09 −0.06
P-07 Skate-Tracker A skateboarding progression app: a prerequisite graph of 40+ tricks, RAG-based coaching, and video comparison against pros. Projects 2025 −0.14 −0.12
P-08 setos.ai Startup Prototype & Pitch A startup prototype that turns a plain-language question into a reading path through more than a million research papers. Projects 2025 0.05 −0.56
P-09 ForgeQB An adaptive quizbowl study app with 500+ users, recommending from 160,000+ questions clustered into micro-categories. Projects 2025 −0.54 −0.10
P-10 Inter-Agent Misalignment: Probing & Activation Steering Testing whether a misaligned coding sub-agent can pull its conductor LLM toward reward hacking, and steering it back through its activations. Projects 2026 0.52 −0.74
W-01 ForgeQB: Recommending Quizbowl Questions How we serve question recommendations for a trivia vertical where both timing and accuracy matter, when every question comes with only a broad category and a broad difficulty. Writing 2026 −0.64 −0.05
W-02 Thoughts on America's 250th An exploration of my relationship with the USA and its founding ideals. Writing 2026 −0.68 0.13
W-03 Dreams or Delusions An exploration asking whether naive dreaming serves us better than uncomfortable truth. Writing 2025 −0.29 0.23
H-01 Badminton The fastest of the racquet sports: short rallies, a lot of footwork, and very little time to think. Hobbies — −0.85 0.09
H-02 Pickleball The slower racquet sport, where placement and patience beat power. Hobbies — −0.76 0.09
H-03 Thrifting Digging through secondhand racks for the piece nobody else noticed. Hobbies — −0.73 −0.14
H-04 Fashion Clothes as a design problem: proportion, color, texture, and how the pieces work together. Hobbies — −0.47 −0.03
H-05 Art history The stories behind paintings and the movements around them. Hobbies — −0.76 0.17
H-06 Classic literature Older books that still argue with the present, Steinbeck and Camus among them. Hobbies — −0.86 0.27
H-07 Philosophy From Plato to Camus to Kant: the questions underneath the essays on the writing plate. Hobbies — −0.66 −0.00
T-01 Time series Learning from sequences that unfold in time: market returns, disease counts, ICU stays and vibration signals. Technical Interests — 0.39 0.82
T-02 Formal verification Proving that code meets its specification, and using LLMs to write the specifications in the first place. Technical Interests — 0.21 −0.74
T-03 Neural operators Networks that learn maps between functions rather than between fixed-size vectors. Technical Interests — 0.72 0.52
T-04 Physics-ML Machine learning for physical systems, where the signal has structure a generic model tends to miss. Technical Interests — 0.64 0.71
T-05 Representation learning What a model's internal space captures, and whether that space is useful for anything downstream. Technical Interests — 0.64 0.62
T-06 Mechanistic interpretability Reading a model's behavior off its activations, and steering it by editing them. Technical Interests — 0.44 0.05
T-07 Multi-agent safety How agents that work together can pull each other off course, and how to catch it when they do. Technical Interests — 0.56 −0.71
T-08 Blockchain Systems where the rules live on a shared ledger: smart contracts, on-chain commitments and incentives that settle without a middleman. Technical Interests — 0.00 −0.33
T-09 Automated market makers Exchanges that price assets with a formula over a liquidity pool instead of matching buyers and sellers in an order book. Technical Interests — 0.18 −0.44
T-10 Agent evals Measuring what AI agents actually do: benchmarks, harnesses and scoring that hold up when you look closely. Technical Interests — 0.64 −0.55

Each category is printed as its own ink on the same axes. Lift a layer to see what's underneath; where layers overlap, the inks mix.

Ink layers · 6 of 6 on
PlotList