Skip to content
View Tushar-9802's full-sized avatar

Block or report Tushar-9802

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.

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
Tushar-9802/README.md

Hi, I'm Tushar 👋

I'm an ML engineer who fine-tunes small language models on consumer hardware — and builds the products around them. Most of my work sits at the messy intersection of applied research and things people actually use: clinical AI for frontline health workers, offline-first LLM pipelines, and cross-modal transfer learning.

My paper on cross-modal LoRA degradation was presented at IMPACT 2025 (Springer). A second one, on preventing catastrophic forgetting, is in preparation.

Systems Dev intern @ Cognavi  ·  Uttar Pradesh, India  ·  Portfolio


Research

Cross-Modal Transfer Learning in Domain-Adaptive Video Summarization IMPACT 2025, Springer — presented Dec 6, 2025 · proceedings forthcoming

I found that LoRA fine-tuning on academic papers degrades video summarization by 14–50%, then built a 5-method curriculum framework to fix it.

System Video ROUGE-1 Video ROUGE-2 CMC
Baseline (zero-shot) 0.263 0.032 0.356
IMPACT 2025 naive LoRA 0.272 0.060
Full framework (Phase 3) 0.417 (+58%) 0.119 (+272%) 0.531

p < 0.005, paired t-test · bootstrap 95% CI · 1,400-line eval harness

Second paper in preparation: Preventing Catastrophic Forgetting in Cross-Modal Summarization.


Projects

Clinical AI

Project What it does Highlights
Sakhi Offline-first Hindi voice-to-form for India's 1M+ ASHA health workers — home-visit conversations become structured government forms with real-time maternal & child referral decisions. 87% end-to-end pass on synthetic Hindi audio · dual deploy: health-center (~15s) + on-device Android via Cactus SDK (~5min) · 6-layer anti-hallucination pipeline · Gemma 4 Good Hackathon
MedScribe Speech → concise SOAP notes in physician shorthand, fully offline on 16GB VRAM. Built for the Google MedGemma Impact Challenge 2026. 16 → 5 min per encounter · 0% hallucinated findings · 90/100 quality · LoRA-tuned on 712 samples for $1.28

Generative AI

Project What it does Highlights
YouTube Transcript Summariser Domain-adaptive summarization for 90-minute lectures on a single 8GB GPU. The negative result that started the research. Published — IMPACT 2025 (Springer) · documented a 14–50% cross-modal ROUGE drop, p<0.001
Hybrid Dataset Summariser The follow-up that fixes it — a 3-phase curriculum (LoRA+, OPLoRA, EWC, CrossCLR) on Mistral-7B over a 4,324-sample cross-modal dataset. Video ROUGE-1 0.417 (+58%), ROUGE-2 +272% · forgetting held at 14.1% vs 31.4% · all gains p<0.005 · HF model · dataset
Bawarchi Multimodal recipe generation: vision detects ingredients, embeddings + PMI suggest cuisine-appropriate swaps, a fine-tuned LLM writes the recipe. YOLOv8m 66.5% mAP · substitution engine 85% precision · Llama 3.2 3B tuned on 335K recipes (74.1% accuracy)

Data Products

Project What it does Highlights
DPDPA Compliance Legal-tech automation for India's Digital Personal Data Protection Act, 2023 — 50+ regulatory requirements assessed in minutes instead of days. 40 hours → 5 minutes (480× the manual baseline) · ~95% PDF extraction accuracy · 9-table relational DB · 15-question assessment engine · targets the May 2027 deadline
Tech Layoffs Dashboard Interactive analytics on global tech-sector layoffs, 2020–2025, with six custom metrics (efficiency, instability, severity, and more). Live Streamlit · YoY/QoQ trend analysis · company-level deep-dives · Kaggle EDA

Open Source

llmclean — zero-dependency Python library for cleaning and normalizing raw LLM output · PyPI

Model Adapters — trained LoRA adapters for MedScribe, Bawarchi, and the Hybrid Summariser · Hugging Face

Datasets — curated training datasets for all projects · Kaggle


Stack

Python C++ SQL PyTorch Hugging Face CUDA FastAPI Streamlit Pandas NumPy Plotly Jupyter Git

Beyond the badges: Transformers · PEFT / LoRA · QLoRA · Unsloth · bitsandbytes · OPLoRA · EWC · Whisper · YOLOv8 · Mistral · Gemma · Llama 3.2


GitHub


Links

Portfolio · Hugging Face · LinkedIn · Kaggle · tusharbrisingr9802@gmail.com

Pinned Loading

  1. MedScribe MedScribe Public

    Offline clinical SOAP notes in physician shorthand — MedGemma Impact Challenge 2026

    Python

  2. YouTube-Transcript-Summarizer YouTube-Transcript-Summarizer Public

    Code for our IMPACT 2025 paper (Springer, presented Dec 2025; proceedings forthcoming). Rigorous negative result: domain-specific LoRA on 25K academic papers degrades YouTube video summarization by…

    Python

  3. Hybrid-Dataset-Summariser Hybrid-Dataset-Summariser Public

    Cross-modal learning framework preventing catastrophic forgetting in video summarization. OPLoRA orthogonal projection + LoRA+ asymmetric LRs + EWC + curriculum phasing. Video R-1 +58%, R-2 +272% o…

    Python

  4. Sakhi Sakhi Public

    Offline Hindi voice-to-form for India's 1M+ ASHA health workers. Whisper ASR + Gemma 4 (E2B on-device / E4B server) extracts MCTS forms from home-visit conversations + flags preeclampsia/PPH/neonat…

    Python

  5. llmclean llmclean Public

    Small Python library for cleaning the noise out of raw LLM output

    Python 4

  6. hindi-normalize hindi-normalize Public

    Convert Hindi number words to digits and normalize Devanagari ASR/OCR text — zero-dependency

    Python