Skip to content
View riya0920's full-sized avatar

Block or report riya0920

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.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

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

Report abuse
riya0920/README.md

Riya Soni

ML/AI engineer. I build inference systems that hold a latency budget under real load, and evaluation harnesses that can fail a build.

MS Computer Science, Stevens Institute of Technology (May 2026). Currently Data Analyst at Giraffe Media Group, building SMS analytics infrastructure on Redshift and shipping internal AI tooling.

Open to ML/AI Engineer roles starting May 2026. New York metro.


Featured

streaming-tts-serving VITS behind a decoupled Triton backend written in C++, TensorRT FP16 engines under it, and a Go gateway holding the WebSockets. Latency stops tracking sentence length: 3,200 held sessions across 2 GPUs at 113.8 ms p99 time-to-first-audio, 0 underruns and 0 rejections, 350x aggregate real-time factor. Demo with audio · Triton TensorRT C++ Go CUDA

asr-serving-vllm-k8s Whisper-Large on vLLM, autoscaled on Kubernetes by queue depth rather than GPU utilization (KEDA + Prometheus), shipped through Argo CD behind a word-error-rate gate in CI. 21.8x throughput over a sequential HF baseline, 1.60% WER reproduced to four decimals across three hosts, 92% lower cost per audio hour. Run it in Colab · vLLM Kubernetes KEDA Argo CD Prometheus

rag-eval-harness Hybrid retrieval (BM25 + dense, RRF fusion) written without a framework, a hand-authored golden set, and a CI gate that blocks a retrieval regression, plus an inverted CI step that fails the build if the gate itself stops catching a planted regression. docs/LIMITATIONS.md states in writing that this project's own headline comparison sits inside the noise band. Evaluation results · Python BM25 RRF GitHub Actions

profiling-training-study An optimisation ladder where every rung changes exactly one thing and the profile picks the order, not a blog post. Run on CPU, then on a rented NVIDIA L4: one rung (workers) is +298% and the eight rungs after it sum to +17%, while data-wait never falls below 79%, so the GPU stays starved through every kernel-level tweak. bf16 flips sign between the two machines and is still reported as unestablished, because the spread is 18% wide. Negative results stay in the table. Measured results · PyTorch DDP CUDA torch.compile

self-healing-rl-pipeline Four agents over MCP and A2A (Monitor, Diagnostics, Repair, Verification) that detect drift in a DQN recommender, diagnose the root cause, retrain the policy, and verify the fix with no human in the loop. PyTorch FastAPI MCP A2A LangSmith

predictive-maintenance-rul-platform Remaining-useful-life prediction on real NASA C-MAPSS data, all four sub-datasets: PHM08 asymmetric scoring, piecewise RUL cap, alpha-lambda accuracy, an alarm policy priced on lead time against nuisance-alarm cost, an ONNX edge benchmark, and a drift-triggered retrain rule. docs/RESULTS.md is written by the training run, not ahead of it. PyTorch ONNX scikit-learn


Research

Mechanistic interpretability of decoder-only transformers, advised by Matthew Finch (Senior ML Engineer, Bloomberg LP). I measure polysemanticity across GPT-2 and Pythia attention heads and prune parasitic SVD components post-hoc, with no retraining. Eleven notebooks, each standalone.

The headline pruning result is currently in-sample: greedy selection and reporting run over the same corpus, and the effect sizes sit near the per-document noise floor. Held-out validation is the next step, and I would rather write that here than quote the number.


Stack

Python PyTorch C++ Go TypeScript · Triton Inference Server TensorRT vLLM ONNX · Kubernetes KEDA Argo CD Docker GitHub Actions · LangGraph MCP ChromaDB FastAPI · SQL (Redshift, PostgreSQL, DuckDB) dbt Kafka Prometheus Grafana AWS


Reach me

LinkedIn · Email · New York, NY

Pinned Loading

  1. streaming-tts-serving streaming-tts-serving Public

    Streaming TTS at 113.8 ms p99 time-to-first-audio: VITS on a decoupled C++ Triton backend with TensorRT FP16 engines and a Go WebSocket gateway. 3,200 held sessions across 2 GPUs, zero underruns.

    Python

  2. asr-serving-vllm-k8s asr-serving-vllm-k8s Public

    Whisper-Large on vLLM, autoscaled on Kubernetes by queue depth (KEDA) and shipped by Argo CD behind a word-error-rate gate in CI. 21.8x throughput over a sequential HF baseline, 1.60% WER.

    Python

  3. rag-eval-harness rag-eval-harness Public

    Hybrid retrieval without a framework, a 35-example golden set, and a CI eval gate verified by a planted regression.

    Python

  4. profiling-training-study profiling-training-study Public

    Profiling-driven training optimisation on CPU and a rented NVIDIA L4: one rung is +298% and the eight after it are noise, the GPU stays data-starved through every kernel tweak, and negative results…

    Python

  5. predictive-maintenance-rul-platform predictive-maintenance-rul-platform Public

    RUL prediction on real NASA C-MAPSS (all four sub-datasets): PHM08 asymmetric scoring, piecewise RUL cap, alpha-lambda, alarm policy with lead time and nuisance-alarm economics, ONNX edge benchmark

    Python

  6. self-healing-rl-pipeline self-healing-rl-pipeline Public

    Autonomous self-healing recommendation system. Multi-agent A2A architecture with Monitor, Diagnostics, Repair, and Verification agents that detect drift, diagnose root cause, and retrain a DQN poli…

    Python