Skip to content
View AmitAK1's full-sized avatar

Achievements

Achievement: YOLOAchievement: Pair Extraordinaire

Achievements

Achievement: YOLOAchievement: Pair Extraordinaire

Highlights

  • Pro

Block or report AmitAK1

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
AmitAK1/README.md

Amit Anil Kamble | AI/ML Engineer

Applied Machine Learning • Deep Learning • Computer Vision • NLP

I build production-minded ML and CV systems end-to-end—from data + modeling to evaluation, explainability, and deployment-ready pipelines.

  • Core focus: CV (detection/segmentation), NLP/Transformers, medical imaging, efficient fine-tuning
  • Tooling: PyTorch/TensorFlow, Hugging Face, OpenCV, FastAPI, Docker, Git
  • Looking for: AI/ML / CV / NLP internships & entry roles

⭐ Recruiter Highlights (Latest AI/ML-Ranked Projects)

1) Darukaa Eco Scientist

Environmental AI scientist with hybrid RAG + causal reasoning

  • Latest project (updated Sep 18, 2026) aligned to real-world AI/ML problem-solving
  • Hybrid retrieval with dense embeddings + BM25 + Reciprocal Rank Fusion (RRF)
  • Structured reasoning over FAO/IPCC data with deterministic validation and explainable outputs

Live / Demo: https://darukaa-eco-scientist-assignment.streamlit.app/

Tech: RAG, LanceDB, BM25/Tantivy, Sentence Transformers, Gemini, Streamlit

2) Text-Guided BraTS Segmentation

Multimodal medical imaging: text-guided brain tumor segmentation (BraTS)

  • Newest research direction for multimodal CV + medical AI
  • Positioning: bridges natural language guidance with 3D/medical segmentation workflows

Tech: Medical imaging, segmentation, multimodal learning

3) Automated Bone Age Estimation

Pediatric bone age prediction from hand X-rays (RSNA)

  • R² = 0.9169 | MAE = 9.04 months on 12,600+ images
  • Transfer learning with Xception + regression head
  • Added Grad-CAM explainability and checked gender bias

Tech: TensorFlow/Keras, transfer learning, Grad-CAM, medical imaging

4) SanitiSense AI

Civic AI operating system for urban sanitation: multimodal reporting + route optimization + epidemic risk forecasting

  • Hackathon-grade, end-to-end system design with strong real-world product thinking (citizens → workers → authorities)
  • Multimodal inputs: photo-first reporting + multilingual voice notes; spam filtering & verification workflow
  • Cloud-native architecture with AWS Lambda + API Gateway + DynamoDB + S3, and GenAI services via Amazon Bedrock

Live / Demo: https://main.d18a1q87hofbl9.amplifyapp.com/

Tech: AWS (Lambda, API Gateway, DynamoDB, S3), Amazon Bedrock, RAG/Knowledge Base, Next.js, TypeScript, Python

5) Real-Time Object Surveillance System

YOLOv8-based object tracking + monitoring

  • Real-time detection/tracking pipeline with OpenCV
  • Practical CV system design: alerts + monitoring use-case

Tech: YOLOv8, OpenCV, Python

🧪 Additional Projects

🛠️ Technical Skills

Languages: Python, C/C++, JavaScript, SQL
ML/DL: PyTorch, TensorFlow/Keras, scikit-learn, Hugging Face Transformers
Computer Vision: OpenCV, YOLO, CNNs, Transfer Learning, Segmentation
NLP: BERT/CodeBERT, LoRA, PEFT, Transformers
Backend/Tools: FastAPI, REST APIs, Docker, Git, Jupyter, Pandas, NumPy

📬 Contact

Pinned Loading

  1. Automated-Bone-Age-Estimation-using-Deep-Learning Automated-Bone-Age-Estimation-using-Deep-Learning Public

    Deep learning system for automated pediatric bone age estimation from hand X-rays using Xception and transfer learning, achieving R² = 0.9169 and MAE = 9.04 months with low gender bias and Grad-CAM…

    Jupyter Notebook 1

  2. Neural-Vulnerability-Scanner Neural-Vulnerability-Scanner Public

    CodeBERT + LoRA fine-tuning for C/C++ vulnerability detection | F1 = 74.3% | PyTorch, HuggingFace Transformers, PEFT

    Jupyter Notebook 2

  3. pyclean pyclean Public

    PyClean DSL is a full-stack compiler that enables data cleaning and EDA using an English-like DSL. It implements Lexer, Recursive Descent Parser, AST, Semantic Analysis, and generates executable Py…

    Python

  4. missing-object-surveillance missing-object-surveillance Public

    Real-time object tracking and monitoring system using YOLOv8

    Python