My Journey

A progression from technical foundations to production-scale AI systems.

This page traces how my work evolved, from modeling and experimentation into data products, distributed systems, and more operator-aware ML....

2017 - 2021

NIT Silchar

B.Tech in Electronics and Communication Engineering

Built my mathematical and systems foundation through coursework in linear algebra, algorithms, and neural networks. This phase shaped the base for my later work in ML, data systems, and AI.

Foundations

Automatic License Plate Recognition

An early computer vision project that sharpened my understanding of robustness under blur, distortion, and low-light conditions.

Early Customer Analytics Projects

Included a purchase propensity model and a clustering workflow, marking a shift toward business-oriented ML and tabular data problems.

2021 - 2022

Wipro

Junior Data Scientist

Moved into product-facing ML through recommendation systems, evaluation logic, APIs, and deployment workflows. This phase pushed my work toward practical ML under real engineering constraints.

Applied ML
2022 - 2024

Samsung Research India

Data Scientist

Worked on production-scale data and ML systems across SmartTV analytics and ad delivery. This is where my work became deeply systems-oriented and production-focused.

Production Systems

Ocular Disease Analysis

A multimodal healthcare project that strengthened my work in transfer learning, preprocessing, and structured evaluation.

2024 - 2026

University of Maryland

M.S. in Data Science

Expanded into graph analytics, reinforcement learning, and intelligent interfaces through coursework and projects. This phase broadened my focus from prediction and analytics to more interactive AI systems.

Advanced AI Systems

ClimaSense

A reinforcement learning system for HVAC optimization that pushed my work from prediction into control and simulation.

Real-time Bitcoin Analytics

A graph analytics system that brought together ingestion, schema design, natural-language querying, and user-facing interfaces.

Current Work

What I’m building now.

These projects reflect where my work is focused now: productized AI, planning systems, and more model-level experimentation.

Planera

An inspectable analytics copilot for structured data that keeps planning, SQL generation, and validation visible.

AI InterfacesAnalytics SystemsOrchestration

Contour

A sprint planning copilot that turns high-level goals into scoped tasks, priorities, and assignments based on team skills and capacity.

Workflow SystemsPlanningProduct Design

Prompt-Adaptive Diffusion Optimizer

An adaptive inference system that reduces SD v1.5 latency with prompt-aware step allocation and latent-convergence early stopping.

Inference SystemsLatency OptimizationDiffusion
Capability Map

Skills that compound in production.

The emphasis here is on strengths that grow together in production: platform thinking, applied ML, decision systems, and disciplined delivery.

Systems
Distributed ETLObservabilityCost EngineeringService Design
Applied ML
Recommendation SystemsForecastingGraph MLComputer Vision
Platforms
AWSKubernetesAirflowPySpark
Delivery
Product ThinkingExperimentationCI/CDOperator Trust