Worked on controllable generation in large language models ( LLMs ).Built Prompt Layer for internal LLM Gateway} project.Led project on a novel reasoning algorithm for LLMs, via classification. Submitted to a prominent conference in NLP.

Worked on developing mathematical pricing and volatility models (and adjacent infrastructure) for linear and nonlinear rates derivatives products.

• Worked on developing the open source QuTiP library.• Worked on developing an open source library to model Open Quantum Systems using HEOM to model the environment as a mixture offermionic and bosonic baths, with Dr. Neill Lambert.• Worked on studying Exceptional Points and exotic properties of large non-hermitian systems using a novel algorithm developed by me, withDr. Fabrizio Minganti.• Worked on optimizing Noisy Mean Field Annealing for NP hard problems using ideas from quantum thermodynamics, with Dr. Neill Lambert,Dr. Clemens Gneiting, and Dr. Ken Funo.• Was invited to IBM’s camp. Worked with Dr. Rudy Raymond from IBM Research to develop a quantum overlapping tomography module usingthe idea of perfect hash families based on Frank Wilczek’s paper. Studied quality of entanglement in IBM’s machines.

• Worked on statistical analysis of returns of several strategies for trading crypto derivatives.• Developed a package for internal use, and conducted deep-dive quantitative work into tracker strategies.• Presented research to institutional investors, including leading hedge fund managers.

• Worked under Dr. Karthik Ramachandra, at Microsoft Research India, Bengaluru on Project Froid• Implemented fix to a part of Froid that improved conditional diamond processing from exponential to linear• Contributed to the open source Ibis repo, implemented an adapter to MSSQL• Worked on PyFroid

• Broadly worked on Matrix Product States and DMRG• Freshly extended and implemented DMRG to Hermitian matrices representing new types of physical systems called Potts Models• Worked on Unitary Time Evolution of Hamiltonians using Trotter Gates• Explored some ideas from quantum computing and information theory

• Developed market-neutral, medium-frequency Alphas that predict future stock returns• Acquired basic knowledge of financial instruments, and investigated and implemented recent academic research• Developed algorithms to filter and combine Alphas

• Was shortlisted by ACM India from a national pool of applicants to attend (top 100)• Attended talks on contemporary cryptology research by, and worked alongside leading Indian cryptographers

• Developed Natural Language Processing modules for an on-device, locally available voice recognition interface• This entailed Morphological Segmentation of text corpora, Coreference Resolution of expressions, Sentiment Analysis of input text and hierarchical representation of concepts to compute similarityUsed PostgreSQL, Scikit-learn, StanfordCoreNLP, Numpy, Pandas, BeautifulSoup, ElasticSearch, NLTK along with some other proprietary services