Rahul Kumar

ML Engineer @ Google Core ML (Keras)

Hi, I'm Rahul. I make frontier GenAI models run on the devices in people's pockets, and I contribute the open-source plumbing that gets them there. If it involves Gemma, Llama, or quantized transformers on NPUs, it probably crosses my desk.

60+ PRs to Google's Keras 1,003 contributions / yr 75% memory ↓ via QAT 10M+ synthetic samples Top 0.05% HackerRank

⚡ Open-Source Impact

108 PRs authored across Google's Keras ecosystem — 43 merged (keras 26 · keras-hub 17) · 35 in review (keras, keras-hub, keras-io, litert-torch)

169

Commits

Pull Requests

Issues

Reviews

1,230

Contributions / yr

keras-team/keras

#22758 — Add LiteRT export support for PyTorch backend merged

Enabled on-device deployment of PyTorch-trained Keras 3 models via TensorFlow Lite / LiteRT.

keras-team/keras-hub

#2132 — Llama 3.1 merged

Implemented Llama 3.1 natively in KerasHub.

keras-team/keras

#22362 — Fix save/load for custom models with sublayers in nested lists merged

Fixed a serialization bug that broke persistence for a whole class of custom architectures.

keras-team/keras

#22998 — Fix torch backend SymInt handling in convert_to_tensor and slice merged

Made symbolic-shape tracing survive the torch export path — required for reliable LiteRT conversion.

keras-team/keras

#23186 — perf(torch): SDPA is_causal dispatch + bounded causal-mask cache in MultiHeadAttention in review

Part of a 13-PR torch-performance series: FlashAttention causal dispatch and capped mask-cache growth.

keras-team/keras-hub

#2705 — Implement KerasHub → LiteRT-LM export with prefill / decode signatures in review

Optimised on-device LLM inference with separate prefill and decode phases (Gemma 3 270M on Android, end-to-end validated).

keras-team/keras-io

#2373 — Guide: Exporting Keras models to LiteRT with PyTorch backend in review

The official keras.io guide for LiteRT export from the PyTorch backend.

🏗 Edge Deployment Architecture

How frontier models travel from research to the NPU in your pocket — via Keras and LiteRT.

flowchart LR A["🤖 Frontier Model\nGemma · Llama · Flux"] --> B["Keras 3\nPyTorch / JAX / TF"] B --> C["LiteRT Export\nPrefill · Decode sigs"] C --> D["On-Device Runtime\nNPU · DSP · GPU"] D --> E["📱 Android / Edge\nGemma 3 270M ✓"] style A fill:#eef2ff,stroke:#4f46e5,color:#1f2328 style B fill:#eef2ff,stroke:#4f46e5,color:#1f2328 style C fill:#dbeafe,stroke:#3b82f6,color:#1f2328 style D fill:#d1fae5,stroke:#10b981,color:#1f2328 style E fill:#d1fae5,stroke:#10b981,color:#1f2328

Keras 3 abstracts the training backend; LiteRT-LM export produces hardware-optimised prefill + decode TFLite graphs. Gemma 3 270M has been validated end-to-end on Android.

💼 Experience

flowchart LR A["Samsung R&D\nJul 2022 – Sep 2024"] --> B["Qualcomm\nSep 2024 – Jan 2025"] B --> C["Google\nFeb 2025 – Present"] style A fill:#f7f8fa,stroke:#d0d7de,color:#1f2328 style B fill:#fff7ed,stroke:#f59e0b,color:#1f2328 style C fill:#eef2ff,stroke:#4f46e5,color:#1f2328

Google — Core ML / Keras Team

Software Engineer, Machine Learning · Feb 2025 – Present · Hyderabad, India

KerasLiteRTPyTorch JAXTensorFlowStarlarkONNX

Shipped LiteRT export for the PyTorch backend in Keras 3, enabling on-device deployment of PyTorch-trained models via TFLite / LiteRT (keras#22758).
Built LiteRT-LM export for KerasHub with prefill / decode signatures — validated Gemma 3 270M on Android, end-to-end.
Implemented Keras-native Llama 3.1 and Multimodal Gemma 3; fixed 4/8-bit quantization instabilities across backends.
Authored the official keras.io guide for LiteRT export (in review).
Migrated Adbrain ad-recommendation model to Keras 3; TF 2.20 / Python 3.13 / Orbax compatibility fixes.
108 PRs authored across keras, keras-hub, keras-io, and litert-torch (43 merged, 35 in review): bug fixes, perf optimisations, cross-backend compatibility.

Qualcomm

Machine Learning Engineer · Sep 2024 – Jan 2025 · Hyderabad, India

QATSnapdragonNPU DSPLiteRTC++Android

Applied Quantization-Aware Training (QAT) on Snapdragon mobile SoCs, reducing memory bandwidth 75% while retaining model quality.
Redesigned Magic Keeper generative inpainting pipeline for the Snapdragon Summit demo — NPU and DSP accelerated.
Built voice-controlled Camera Copilot demo for on-device AI showcase at client briefings.
Delivered hardware-accelerated Android reference apps demonstrating GPU, DSP, and NPU delegation for benchmarking.

Samsung R&D Institute India (SRI-N)

R&D Engineer — Computer Vision · Jan 2022 – Sep 2024 · Noida, India

YOLOMask R-CNNTFLite OpenCVKotlinJavaAndroid

Knox Capture: End-to-end damaged-barcode scanning pipeline — 10M+ synthetic samples via geometric/probabilistic transforms; accuracy 70% → 90% with YOLOv11, Mask R-CNN, U-Net.
SRIN-Satyapan: Anti-cheat proctored exam platform — 2,000+ concurrent users, 10,000+ registered faces, real-time face-recognition auth on Android + web portal. MD Appreciation Award.
Face recognition optimisation: Embedding search O(n) → O(log n); lookup time 1,200 ms → 87 ms on Android tablets serving 10k+ users.
Drop detection: Replaced deep learning with statistical feature engineering — 99% parameter reduction, 95% power reduction, accuracy drop <2%.

🚀 Projects

🧠

Keras / KerasHub — LiteRT Export Pipeline

On-device deployment infrastructure for frontier models (Gemma, Llama, Flux) via Google's Keras 3 ecosystem. Covers PyTorch-backend export, LiteRT-LM prefill/decode signatures, and the official keras.io documentation.

View PRs →

📈

RL Alpha Labs — Algorithmic Trading

JAX-native PPO and DQN agents trading across 746 NSE instruments with 3.5+ years of OHLCV data. 128 vectorised environments for parallel rollout, full portfolio simulation with risk controls.

Live Demo → Private repo · available on request

📄

ATS Optimizer

Automated job-application pipeline: ATS scoring via NLP keyword extraction + semantic similarity, LLM-powered résumé tailoring, multi-portal scraping (LinkedIn, Indeed), and automated form submission. Built for 50+ applications per day.

Code →