Nikhil Solanki
Exploring how software, data, and machine learning can solve real problems.
I'm a software/ML engineer with a background in computer science and mathematics, currently pursuing my M.S. in Computer Science at Georgia Tech. I'm drawn to problems that combine strong engineering with data, machine learning, and real-world product impact.
Recently, I've been working on ML-driven transaction classification and merchant intelligence at an early-stage fintech startup, along with side projects in reinforcement learning, time-series modeling, and applied machine learning. I'm currently looking for software engineering, machine learning, or data-focused roles where I can contribute, learn quickly, and build useful systems.
Projects
Research
Research-style technical reports exploring machine learning through controlled experiments, algorithm comparison, and applied analysis across supervised, unsupervised, optimization, and reinforcement learning settings.
Compared Decision Trees, K-Nearest Neighbors, Support Vector Machines, and Neural Networks across binary and multiclass classification tasks, analyzing model complexity, learning behavior, timing, and class-level performance.
Studied how optimization and regularization choices affect neural network training by comparing randomized optimization algorithms, Adam optimizer ablations, convergence behavior, and generalization under controlled experimental budgets.
Applied clustering and dimensionality reduction techniques to uncover dataset structure, evaluate representation quality, and measure how unsupervised transformations affect downstream neural network performance.
Implemented Value Iteration, Policy Iteration, Q-Learning, and SARSA across stochastic and control-based environments, comparing convergence, policy quality, discretization effects, and learning stability.
Experience
Founding Machine Learning Engineer
MIQA – Lakeland ScientificBuilding MCC Predict, a merchant intelligence system that infers Merchant Category Codes from transaction data, descriptors, and merchant metadata. Working across ML pipelines, feature engineering, data enrichment, LLM-assisted merchant research, and API-facing prediction workflows.
Software Engineer / Systems Engineer Intern
Lockheed MartinDeveloped, deployed, and tested Aegis Combat System software loads for naval systems. Built Python and Robot Framework regression tests, authored Bash utilities for environment setup, and helped investigate real-time software issues on active naval vessels.
Application Developer Intern / Part-Time
The Vanguard GroupWorked on Personal Advising Services and financial planning systems. Refactored large code paths, enhanced advisory microservices, migrated workflows to AWS Lambda and DynamoDB, and built observability dashboards with Splunk and Honeycomb.