Redmond, WA · Data Scientist · Dancer

Abhinava Maddha

I build ML systems at Microsoft by day, practice Kuchipudi on weekends, hunt for good views on trails, and read old books on rainy afternoons. Always looking for interesting people to share ideas with.

Let's connect See my world →
your photo here click to upload
About

Data, dance, and everything in between

I'm a Data Scientist at Microsoft in Redmond, WA — where I build ML attribution models, run NLP projects on Azure's Learn platform, and help 19 million monthly users find what they're looking for. My work lives at the intersection of user behavior, language, and signal.

I hold an MS in Computer Science from UMass Amherst and a BS from BITS Pilani. My research has spanned semantic parsing, EEG signal classification, neural rendering, and deep reinforcement learning.

But I'm equally shaped by the years I spent training in Kuchipudi under Padmasri Dr. Sobha Naidu — a discipline that taught me precision, expression, and patience. I still practice weekly. I grew up singing Carnatic music under G.V. Prabhakar, and while I've stepped back from formal practice, music is woven into how I think.

Right now I'm learning hip-hop and jazz funk, slowly migrating from ukulele to guitar, and always looking for people to jam with. If that's you — reach out.

MS
UMass Amherst Computer Science, 2023–2025
MBA
NYU Starting 2026
BS
BITS Pilani Electronics & Instrumentation, 2018–2022
20+
Years of Kuchipudi Trained by Padmasri Dr. Sobha Naidu
19M
Monthly active users Impacted by dashboards I've built at Microsoft
∞
Books, mountains, and curiosities Currently reading classical literature; currently hiking for views
Experience

What I've built

Jan 2025 — Present
Microsoft
Data Scientist
  • Built Azure Portal engagement dashboards for 19M monthly active users with A/B testing
  • Set benchmarks for AI Q&A metrics; improved model performance by 6%
  • Developing ML attribution models evaluating the impact of Learn on Azure Portal consumption
  • Spearheading NLP projects analyzing sentiment, intent, and user cohorts for personalization
May–Aug 2024
Microsoft
Data Scientist Intern
  • Quantified ROI for Learn usage by Enterprise customers to inform leadership on content investments
Jan–May 2024
AMD
MLSE Intern
  • Generated 3D-convolution configs to improve heuristics for GPU kernel optimization
  • Compared CVAE, CGAN, and SMOTE methods using cosine divergence and KL divergence
Jan–Jul 2022
Uber Eats
Data Analyst Intern
  • Increased merchant acquisition by 18% with automated dashboards for Merchant Universe Segmentation (ANZ)
  • Built quarterly business reports for promos and campaigns
Life

Beyond the Business OKRs

I live seasonally — snowboarding in winter, hiking for panoramic views in summer, paddleboarding on the right afternoons. All year: tennis courts, classical books, and weekend practice sessions in Kuchipudi. Plus: ballet performances, orchestras, and musicals whenever I can find them.

🎭
Kuchipudi
Classically trained in this Indian classical dance form under Padmasri Dr. Sobha Naidu. Still practice weekly — the discipline never leaves you.
🎵
Carnatic Music
Trained vocally under G.V. Prabhakar. A heritage I carry quietly — I still sing to myself even if I don't practice formally anymore.
💃
Hip-Hop & Jazz Funk
Currently in training — a different energy from Kuchipudi but the same love of movement and expression.
🎸
Guitar (in progress)
Making the switch from ukulele. Looking for people to jam with — reach out if that's you.
🏔️
Mountain pursuits
Snowboarding in winter, hiking for views in summer, the occasional paddleboard when the water's right. Seasonal adventures, always.
🎾
Tennis
All year, rain or shine. The one constant across seasons and cities.
📚
Classical Literature
Something about older books — the pace, the depth, the way language was used differently. There's always one on the nightstand.
🩰
Ballet & Performing Arts
Ballets, orchestras, and musicals — a few times a year, but always something to look forward to.
📷
Photography
A visual storyteller at heart — documenting the mountains, the performances, and the quiet moments in between.
💡
Business & Technology
Interested in how things are built, how markets work, and what's changing. Always up for a good conversation about it.
Photography

Through my lens

Add your own photos — click any frame below

Projects

Things I've worked on

NLP · Research
Case-based Reasoning for Semantic Parsing
Graduate research at IESL, UMass Amherst. Aggregated span alignments from tokens to improve generalization in semantic parsing for case-based reasoning.
View →
Deep Learning · Research
Emotion Recognition from Multisensory Physiological Data
CNN model for emotion classification using multisensor physiological signals. Best accuracy of 62%. Research at BITS Pilani.
Report →
Deep Learning · Research
Command Classification using Brainwave EEG Tensors
Evaluated CNN, RNN, and LSTM models for multiclass classification of EEG signal tensors with auditory attention. 24% accuracy improvement over baseline.
In progress
Computer Vision · Research
Neural Rendering in Construction
Experiment design comparing Autoencoders with CNNs (NeRFs) for neural rendering of low-resolution images. Research at Arizona State University.
Slides →
Reinforcement Learning · Research
PID Control using Deep Reinforcement Learning
Automated PID control supervision using Deep RL as an extension to the CACLA algorithm. Research at CSIR-CEERI.
Report →
NLP
Hate Speech Detection in Low-Resource Language Transcripts
Low-resource language learning and domain adaptation using transformer-based cross-lingual models like mBERT with adapters on TV show transcripts.
Report →
Full Stack
EleNa — Elevation-based Navigation
Navigation system for bikers and hikers that optimizes routes based on elevation gain/loss preferences.
GitHub →
NLP
Hate Speech Detection against Women
NLP classifier for identifying hate speech targeting women across online platforms.
GitHub →
Information Retrieval
Anime Recommender System
Recommendation engine built using information retrieval techniques for anime content discovery.
GitHub →
Achievements

A few things I'm proud of

Dance
Kuchipudi Prodigy Gold Medal
Awarded for excellence in Kuchipudi classical dance.
Record
2 Guinness World Record Certificates
Part of two record-setting achievements.
Music
Annual Carnatic Music Award
Recognized for vocal excellence in Carnatic music.
Academia · 2015
Salutatorian
Graduated second in class.
Academia · 2015
Mathematics Excellence Award
Recognized for outstanding achievement in mathematics.
Academia · 2015
Computer Applications Excellence Award
Awarded for exceptional work in computer applications.
Tech · 2014
National Cyber Olympiad
92.24 percentile internationally.
Research · 2021
PS-I Special Scholarship
Awarded for internship excellence at CSIR-CEERI.
Leadership · BITS Pilani
Dance Head, Swaranjali Classical Arts Club
Led and choreographed Kuchipudi performances as head of the classical arts club at BITS Pilani.
Leadership · 2014–2015
High School Sports Vice Captain
Elected by peers to lead the sports program in high school.
Connect

Let's be in touch

If you're into tech, business, dance, classical music, mountains, or just want to chat about ideas over coffee — I'd love to hear from you. Reach out on LinkedIn.

or email me at abhiem.business@gmail.com