Abhiram Dharme

— CSE student at IIT Delhi · graduating 2027

Abhiram Dharme

अभिराम धर्मे

Computer Science student working across deep learning, wearable biosignals, and ML systems.

IIT Delhi·CycloFormer · NeurIPS 2026 under review·model design · data scarcity · reliable engineering
Selected research
2025 —
First author

CycloFormer

A rotation-invariant transformer for wrist-sEMG hand pose estimation, with a variance-corrected scaling law for the data-limited regime.

NeurIPS 2026 (under review)

  • Exact ℤ₁₆ invariance to wristband donning rotation, built from a channel-shared TDS-CNN, circular rotary position embeddings, and permutation-invariant attention pooling.
  • New state of the art on emg2pose: a 4M-parameter model surpasses the strongest prior baseline on every generalization split with 33% fewer parameters.
  • A variance-corrected scaling law for the data-limited regime, fit on a 5×8 model–data grid and explaining 98.8% of the observed variance.
  • First controlled quantitative comparison of sEMG and egocentric vision under fingertip occlusion.
Oct 2025 — May 2026
IIT Delhi contingent

RRM Plus — Client-Aware WiFi Resource Management

A graph-neural-network controller for dense WiFi networks, trained with safe reinforcement learning under operational guardrails. Built for the Arista Networks problem statement at Inter-IIT Tech Meet 14.

Inter-IIT Tech 14 · Arista Networks

  • Generated a 1,000-scenario, 4,000-AP, 24,000-client WiFi training corpus on the Komondor / NS-3 simulation stack after ruling out hardware testbeds and MATLAB simulations for scalability.
  • Modeled the network as a heterogeneous graph (AP, client, interference edges) and trained a GNN over per-AP configuration parameters and per-client signal quality / QoE metrics.
  • Trained the policy with Constrained Policy Optimization + a safety critic — every configuration update is screened through a trust-region check on predicted QoE impact and per-AP change budgets.
  • Designed a multi-timescale controller: a fast loop (second-to-minute steering) with strict guardrails, a slow loop (hourly global stabilization), and an event loop that rolls back any action causing > 10% edge-throughput drop.
all research →
Experience
May 2026 — Jul 2026

Google

Software Engineering Intern · Bangalore

May 2025 — Jul 2025

Coinbase

ML Engineering Intern · CBGPT · Bangalore

Built a knowledge-gap classification pipeline for the CBGPT customer-support assistant — LLM-based topic modeling with a FAISS retrieval layer and batched entailment, surfacing under-documented support areas at scale.

Jan 2024 — Mar 2024

Torch Investments

ML Engineering Intern · New Delhi

LightGBM and CatBoost models for Fortune-500 equity return prediction inside a live $3M+ strategy; 2014–2024 backtest of the pipeline yielded 21% CAGR.

more about me →
Recent writing
May 17, 2026 · 3 min read

When You Wear It in a Different Spot: Symmetry as an Inductive Bias

A wristband doesn't know which way you put it on. CycloFormer doesn't either, by construction.

May 10, 2026 · 3 min read

Scaling Laws Aren't Scaling Laws When You're Out of Data

Chinchilla assumes you have enough data to be in the resolution-limited regime. We didn't. So we wrote one that doesn't.

all writing →