Naman Mawandia

Software Engineer | Android, Systems & Machine Learning

Seeking full-time SWE opportunities focused on building intelligent and scalable applications. Passionate about contributing to the rapid growth of innovative startups.

View Projects View Full Resume

About Me

I am a Software Engineer and a CS MS graduate from SUNY Buffalo with industry experience at Samsung Research. During my tenure at Samsung, I delivered high impact features, from reducing VoIP call latency to under 50ms to deploying deep learning-based video interpolation models for thousands of users.


My expertise lies in building smart mobile applications and reliable backend systems that work seamlessly together. I enjoy the challenge of taking complex technical concepts whether it's an advanced AI model or a networking improvement and turning them into fast, easy-to-use products that can scale to millions of users. I am driven by a passion for "under-the-hood" engineering that results in a noticeably better experience for the person using the app.


C / C++ Python Kotlin Java Go Linux TCP/IP Android TensorFlow

Experience

Software Engineer — Samsung Research

Jan 2022 – Jul 2024

At Samsung Research, I worked across the full mobile stack to deliver intelligent features and improve network reliability. I specialized in integrating on-device Machine Learning, specifically using TensorFlow Lite for real-time video interpolation and automated document organization. On the systems side, I engineered kernel-space C tools to trace TCP packet drops and optimized network diagnostics to ensure ultra-low latency (under 50ms) for VoIP services. My work combined academic rigor, including a co-authored IEEE research paper on D2D communication, with the technical agility required to ship features to millions of global users.

C · Kotlin · Python · Wireshark · Android · TensorFlow

Analyst — Axis Bank

May 2021 – Jul 2021

At Axis Bank, I engineered a suite of analytical tools and performance dashboards to optimize internal business operations. By leveraging predictive modeling and data visualization, I improved RM earnings estimate accuracy by 21% and built gamified simulations that enhanced performance tracking for senior leadership. This role allowed me to apply technical logic specifically K-means clustering and statistical analysis, to solve large-scale business challenges and improve the efficiency of incentive structures within the bank.

Tableau · Excel · K-Means

Projects

Monetrix — Expense Tracker

Engineered a personal finance Android app with real-time spending insights and custom data filtering, achieving a polished user experience for expense and income management.

Kotlin · MVVM · Room · Google Ads · Jetpack Compose Google Play

Research RAG

Designed a research assistant using LangGraph to implement Corrective RAG(CRAG), featuring a self-querying retrieval pipeline, metadata filtering and web-search fallback for arXiv paper analysis.

Python · LangChain · OpenAI · ChromaDB · Tavily GitHub

MeetingScribe

Engineered a voice recording Android app with chunked audio capture, chunk-by-chunk transcription pipeline, and AI summary generation, handling edge cases like call interruptions, headset switching, and process death recovery.

Kotlin · Compose · MVVM · Hilt · Room · OpenAI Whisper GitHub

AI Frame Interpolation

Built a deep learning pipeline to upscale video frame rates from 60 to 120 FPS, utilizing a cloud-native Kubernetes service on AWS for real-time video smoothing.

Python · TensorFlow · CNNs · U-Net · AWS SageMaker · Docker · Kubernetes

App Recommender

Developed a predictive Android widget that suggests apps based on usage patterns, increasing user engagement by 23% through context-aware on-device recommendations.

Kotlin · Cosine Similarity · WorkManager · Broadcast Receivers · Alarm Manager GitHub

16-Bit Processor Design

Designed and implemented a full hardware-software co-design of a 16-bit processor, successfully validated through live FPGA demos for real-world system bring-up.

Verilog · FPGA · Vivado · C/C++ · ALU Design GitHub

Whatsup Messaging App

Built a real-time one-to-one messaging platform with secure email registration and profile management, focusing on streamlined user communication and instant message delivery.

Java · Android Studio · XML · Firebase (Realtime DB & Auth) GitHub

Pintos Operating System

Implemented core kernel features including advanced thread scheduling and system call handlers, significantly improving CPU efficiency by eliminating busy-waiting.

C · OS Internals · Kernel · Thread Synchronization

Spatial Temporal Perception

Implemented a high-performance action recognition arch for driver monitoring, achieving greater accuracy on the DriveAct dataset while maintaining low-latency.

Python · PyTorch · VideoMAEv2 · GCN · Transformers Notebook

Client-Server Chat App

Built a robust Linux-based messaging system using socket programming to manage concurrent connections and offline message delivery without the overhead of multithreading.

C · TCP/IP · Linux GitHub