"why dont you take a crack at it?"
These simple words from my undergrad professor, though I didn't grasp it then, became the unexpected catalyst for my career journey. They sparked a realization in 2018: I was genuinely drawn to research and building products. This spark led me to dive into hackathons with classmates, where we turned ideas into working prototypes.
Fuelled by this early experience, I sought out deeper research understanding, securing an internship at IISc's Robert Bosch Center for Cyber-Physical Systems. This experience was pivotal, giving me a structured approach to dissecting problems and engineering effective solutions.
After completing my Computer Science Engineering in Bangalore, I pursued a less conventional path, driven by a passion for dissecting complex systems, developing innovative solutions, and collaborating with people.
This led me to join an early-stage AI startup immediately after graduation. Mentored by the CEO and the CIO, I quickly honed my AI expertise, becoming the youngest founding member and establishing myself as their in-house NLP specialist. Progressing to Senior Engineer within the team, I tackled challenging technical problems collaboratively -- finetuning SLMs, designing ICL paradigms for LLMs, developing RAGs, large scale text clustering, knowledge graph pipelines, few shot classification, entity recognition algorithms, etc.
This experience, over half a decade, instilled in me the ability to deliver under pressure, employing creativity, innovation, and practical problem-solving. Critically, this environment also presented the opportunity to contribute to pioneering new algorithms and patenting novel AI systems - I now bring these skills with me at Apple, where I work as a Senior Machine Learning Engineer
I also actively advocate for open research and mentoring -- delivering guest lectures at my alma mater, contributing to engineering blogs, judging/mentoring hackathons, etc.
I am now giving myself wings through academic nuance -- taking up the mantle to complete my Masters degree, focusing on applied AI @ NUS, Singapore.
Mentoring at the UNESCO India Africa Hackathon
Making waves in the media - 2019
AI Talks at NITK
Mentoring the next generation to become national winners!
UIA Winners Ceremony -- graced by the Vice President of India!
I am also an applied ML researcher, and I love building innovative solutions
Purple : A novel approach to extreme text classification. Classify between than 3000+ classes, with as few as 30 samples per class. Full Technical Report coming soon!
Blitz Clustering : An unsupervised clustering algorithm to cluster texts. Unlike other algorithms, it has extreme interpretability using only min_size and min_sim as cluster-level arguments! Its also called blitz because its the fastest text-clustering algorithm*
Recursive Expert Delegation : A semi-supervised algorithm that uses Active Learning with LLM agents. I built it because data tagging is tedious...
Graph Overlap based Joint Optimization : A framework to 'stack' multiple embedding models and find more interesting insights via graphs, developed during my master's degree.
LUKE : A novel quantification algorithm to get fine-grained scores of 'how' intensely depicted a particular sentence's emotion is **
* : as measured and compared to HDBSCAN, DBSCAN, OPTICS at high dimensionality
** : patent pending