ML Engineer building production systems at the intersection of research and engineering.
I've worked across gaming, VR, aerospace, medical imaging, and cloud infrastructure — and found that machine learning is where all of it comes together. My work focuses on making ML systems that don't just perform well in notebooks, but hold up in production.
Currently at Arkatechture building financial data infrastructure, and pursuing an MS in Data Analytics at McDaniel College. Previously at Amazon Web Services (Computer Vision Data Platform) and Quantil/CDNetworks.
- 📄 ML research under review at ACM PODS — automated SQL data type inference for enterprise data pipelines
- 🔍 Anomaly detection and time-series forecasting for data ingestion workflows
- ✍️ Writing about production ML and the fundamentals that actually matter → medium.com/@anushreedas.2710
| Project | What it does | Stack |
|---|---|---|
| BERT Confusion Emotion Detection | Fine-tuned BERT on time-series transcribed speech to detect multi-level human confusion with statistical significance testing across model variants | PyTorch, HuggingFace, BERT |
| Handwritten Math Expression Recognizer | Multi-stage pipeline for stroke segmentation, symbol classification, and Symbol Layout Tree generation outputting LaTeX and MathML | SVM, Random Forest, Python |
| CNN Dishwasher-Safe Classifier | End-to-end image classification with transfer learning, class imbalance handling, and per-class evaluation | PyTorch, ResNet, VGG16 |
ML & AI scikit-learn · PyTorch · TensorFlow · BERT/Transformers
Random Forest · ensemble methods · computer vision · NLP
feature engineering · causal inference
Data Engineering ETL/ELT pipelines · Snowflake · AWS (S3, Lambda, Glue)
PySpark · SQL · data quality monitoring · schema management
MS Data Analytics McDaniel College 2025 – 2026
MS Computer Science Rochester Institute of Technology 2019 – 2021
BS Information Technology University of Mumbai 2017
AWS Certified Cloud Practitioner (Dec 2024 – Dec 2027)
I write about what I'm actually building — production ML systems, data pipelines, and why the fundamentals matter before chasing the hype.
- Stop Guessing About Your Image Dataset: A Practical EDA Guide with Pandas and Matplotlib
- Finding Structure in Unstructured Image Dataset with VGG16 + KMeans
