Prev - Engineering @ Deloitte | IBM Qiskit Advocate | Independent Researcher

github

medium

projects

Independent Researcher & Software Engineer

Soumyadip Sarkar • Peerlist Profile

Clustor is a high-performance clustering toolkit that leverages the power of Rust to deliver fast and efficient clustering algorithms. Designed with a Python-first API, Clustor integrates seamlessly into production data pipelines, offering minimal-dependency implementations of common clustering algorithms. Built on PyO3 and maturin, it supports a range of unsupervised learning primitives, making it an ideal choice for data scientists and developers working with large datasets.<h3>Highlights</h3><ul><li>KMeans and MiniBatchKMeans with KMeans++ initialization, Euclidean/Cosine metrics, and streaming partial_fit.</li><li>BisectingKMeans for hierarchical clustering.</li><li>DBSCAN with noise labeling and core sample indices.</li><li>OPTICS for reachability ordering and core distances.</li><li>Affinity Propagation for exemplar-based clustering.</li><li>BIRCH for streaming clustering using CF-tree summaries.</li><li>GaussianMixture with model selection metrics like AIC and BIC.</li><li>Cluster validation metrics: Silhouette, Calinski–Harabasz, and Davies–Bouldin.</li></ul>Clustor is actively developed, with APIs subject to change before the 1.0 release. It is open source, licensed under the Apache License, and available on <a target="_blank" rel="noopener noreferrer" class="break-all link-break text-gray-500 hover:underline" href="https://github.com/alphavelocity/clustor">GitHub</a>.

clustor

Qiskit Inspect offers production-ready debugging, tracing, and analysis tools for Qiskit, enhancing the capabilities of Qiskit v2.0+ with practical features. This tool is designed to assist practitioners in debugging classical control flow, validating sampler output against exact statevectors, and ensuring robust tests for production quantum applications.<h3>Highlights</h3><ul><li>Deterministic Debugging: Step through instructions with full statevector snapshots, classical bit snapshots, and support for control-flow constructs such as IfElseOp, loops, and SwitchCaseOp.</li><li>Prefix Analytics: Compute exact probabilities, sampler-backed probabilities, counts, marginals, expectation values, Shannon entropies, divergences, and total-variation distances for every prefix of a circuit.</li><li>Testing Confidence: Assertion helpers compare counts, probabilities, and statevectors with configurable tolerances for noisy hardware.</li><li>Operational Insights: CSV/JSON export, pandas integration, ASCII/Matplotlib histograms, and structured trace logging make it simple to observe and share what your primitives are doing.</li></ul>Qiskit Inspect is open source and licensed under the Apache License 2.0, making it a valuable resource for developers working with quantum circuits.

Qiskit Inspect

<h3>A Lightweight, High-Performance Mathematics &amp; Quantitative Finance Toolkit</h3>DervFlow is a comprehensive toolkit designed for high-performance mathematics and quantitative finance applications. It offers a range of features including production-ready option pricers, risk analytics, portfolio construction utilities, time-series diagnostics, and yield-curve analytics. These capabilities are backed by rigorously tested numerical kernels implemented in Rust for enhanced safety and speed.Highlights:<ul><li>Options Pricing: Includes Black-Scholes-Merton analytics, implied volatility solvers, and exotic payoffs exposed through ergonomic Python classes.</li><li>Risk Analytics: Offers portfolio Greeks aggregation, historical/parametric/Monte Carlo VaR and CVaR, drawdown analytics, and performance ratios.</li><li>Portfolio Construction: Provides mean-variance optimization, efficient frontiers, and Black–Litterman blending of market and investor views.</li><li>Yield Curves: Features bootstrapping from bonds or swaps, Nelson–Siegel parametrisations, and bond analytics.</li><li>Time Series: Supports return transformations, rolling statistics, autocorrelation diagnostics, and GARCH-family volatility models.</li><li>Monte Carlo Simulation: High-throughput stochastic process simulators with optional parallel path generation.</li></ul>

DervFlow

Minitensor is a lightweight, high-performance tensor operations and automatic differentiation library designed for developers seeking a PyTorch-like experience with enhanced speed and efficiency. Built with a Rust backend, it ensures maximum speed and memory efficiency, while its Python bindings provide a familiar API for easy adoption. Minitensor supports complete neural network layers and optimizers, offering seamless interoperability with NumPy arrays. Key features include automatic differentiation for gradient computation, a modular design for customization, and a high-performance engine optimized with SIMD. The library is currently in active development, making it a cutting-edge choice for developers looking to build and deploy intelligent applications.

MiniTensor

Interactive semantic maps (WordNet) via CLI and web. WordAtlas builds interactive semantic maps from <a target="_blank" rel="nofollow" class="break-all link-break text-gray-500 hover:underline" href="https://wordnet.princeton.edu/">WordNet</a>.

WordAtlas

Test your reflexes and number recognition skills by popping the right bubbles before time runs out!

Bubble Game

A small Rust CLI/library that demonstrates RSA primitives for learning, and provides secure flows using OAEP (encryption), PSS (signing), and a hybrid RSA+AES-GCM scheme for large files.

cryptrsa

Easily design, simulate, and visualize quantum circuits with intuitive tools that allow you to experiment with quantum gates and qubit states effortlessly.