Abhishek
Adinarayanappa

Software engineer and CS grad student at NYU, with a strong background in cloud computing and backend systems. Previously worked at Intel building scalable production software, gaining deep expertise in reliability and system design.

Currently focused on bridging machine learning and software engineering, with a deep interest in building end-to-end systems that bring AI capabilities into real-world software.

Get in touch View resume
Abhishek Adinarayanappa

Education

New York University
Brooklyn, NY
Master of Science, Computer Science Sept 2024 – May 2026
GPA: 3.93 / 4.0
Design & Analysis of Algorithms · Machine Learning · Big Data · Deep Learning · Cloud Computing · Computer Vision
Official Transcript
PES University
Bengaluru, India
Bachelor of Technology, Computer Science & Engineering Aug 2018 – May 2022
GPA: 8.62 / 10.0
Data Structures · Operating Systems · Computer Networks · Linear Algebra · Introduction to Data Science · Database Management
Official Transcript

Work Experience

June 2025
August 2025
Backend Engineering Intern
Marketeq Digital Inc. · Miami, FL
  • Engineered an autocomplete backend in NestJS with RESTful APIs, integrating LocationIQ API with debouncing to reduce redundant API calls by 70% on average, delivering real-time location suggestions.
  • Built a MongoDB-driven job title validation system with 250+ curated entries for user onboarding, leveraging all-MiniLM-L6-v2 sentence transformers for semantic role matching, integrated with profanity filtering and spell-checks.
  • Enhanced the platform's frontend by developing reusable React components and optimizing backend APIs, resolving 20+ P1 bugs.
July 2022
July 2024
Software Development Engineer
Intel Corporation · Bengaluru
  • Containerized 6+ production microservices for Intel® Edge Insights for Industrial using Docker, Kubernetes & Helm, enabling independent builds & deployments to cut build time from 30 to 5 mins. Achieved 96% code coverage via pytest.
  • Provisioned Edge Insights for Industrial stack on Intel® Developer Cloud within 2 months, validating cloud-native compatibility for enterprise use and earning the Intel Department Recognition Award for cross-team collaboration.
  • Built a portable storage abstraction layer for Intel® oneAPI to optimize ML training pipelines and resolving storage-compute bottlenecks, improving ResNet-50 training throughput by 40% on Intel GPUs through efficient multithreading.
  • Engineered secure cloud infrastructure on AWS and Azure for Intel® Trust Authority, a SaaS-based remote attestation platform, using Terraform & Ansible. Automated CI/CD pipelines and environment refresh workflows achieving a 99.95% uptime.
  • Automated end-to-end environment refresh and tenant attestation with GitHub Actions, reducing tenant onboarding time from 10 minutes to 3 minutes.
January 2022
June 2022
Software Engineering Intern
Intel Corporation · Bengaluru
  • Implemented distributed tracing across 3 core microservices (Video Ingestion, Video Analytics, Visualizer) of an industrial edge analytics platform using OpenTelemetry & Jaeger, capturing 5K+ traces/min to enhance system observability.
  • Enhanced error reporting and logging in the platform, achieving a 20% reduction in issue resolution time.
  • Integrated Prometheus to enable real-time monitoring of 1,000+ metrics/min, providing visibility into service health and performance.
August 2021
November 2021
Data Engineering Intern
Kreditbee · Bengaluru
  • Streamlined monthly Credit Loss report generation using Python and pandas on MySQL data, achieving an 80% reduction in manual effort.
  • Configured Lyft Amundsen data catalog on AWS for Snowflake metadata ingestion, indexing 500+ schemas for data discovery.

Teaching Experience

September 2025
Present
Graduate Teaching Assistant — Cloud Computing & Big Data (CS-GY 9223)
New York University · Brooklyn, NY
  • Led in-class technical demos on AWS for a class of 70+ students, covering cloud infrastructure and distributed systems in real-world industry applications.
  • Mentored 12+ graduate students on their final projects, providing guidance on system design, architecture, and technical implementation.
  • Held weekly office hours to help students navigate assignments and understand complex course material.
  • Helped design and grade assignments and exams on cloud-native technologies and industry-standard tools.
September 2025
December 2025
Graduate Teaching Assistant — Big Data (CS-GY 6513)
New York University · Brooklyn, NY
  • Led weekly lab sessions for 35+ students on hands-on data processing and big data frameworks applied to real-world problems.
  • Assisted students during office hours with assignments, course projects, and general technical questions.
  • Contributed to developing assignments and grading rubrics ensuring clear and fair evaluation standards.
  • Assisted in final project presentations and evaluations, assessing quality and technical implementation of student work.

Technical Skills

Languages
Python C C++ Go Java JavaScript TypeScript HTML CSS Solidity Bash
Frameworks & Libraries
PyTorch TensorFlow OpenCV React Django Node.js NestJS NumPy Pandas Scikit-learn REST API
Cloud & DevOps
AWS Azure GCP Docker Kubernetes Helm Terraform Ansible Prometheus Jenkins Git
Database & Data Engineering
MySQL PostgreSQL Elasticsearch DynamoDB MongoDB Redis Hadoop Spark Kafka

Projects

Big Data & AI
DigiPulse
Distributed ML Pipeline for Real-Time Trend Analysis
End-to-end streaming pipeline that ingests BlueSky posts, clusters trending topics with Spark ML, and auto-generates summaries with Gemini in real time.
KafkaSpark MLMongoDBElasticsearchGemini
1M+
Posts processed
Real-time
Topic summaries
Cloud & AI
GitGrep
Cloud-Native Vulnerability Scanner
Serverless platform that scans GitHub repos for security vulnerabilities using Semgrep and delivers AI-powered remediation suggestions via Gemini.
AWS LambdaDynamoDBSQSGemini AISemgrep
<1 sec
Latency
5+
Languages scanned
Cloud & AI
Natural Language Photo Album
Natural Language Image Search
Serverless photo search app that accepts natural language queries to find images by people, objects, and landmarks — powered by Amazon Rekognition for visual analysis and Amazon Lex for intent recognition.
AWS LexRekognitionElasticsearchCloudFormation
NLP
Query interface
CI/CD
Auto-deployed
Big Data
Centralized Scheduling Framework
Yet Another Centralized Scheduler
Custom Hadoop 1.0 scheduling framework featuring 3 different load balancing algorithms (FCFS, Round Robin, and Least Loaded) with master-slave design and socket-based worker communication.
PythonHadoopSocket ProgrammingMultithreading
3
Scheduling algos
Master-slave
Architecture
Cloud
Dining Concierge
Restaurant Recommendation Chatbot
Serverless chatbot on AWS Lex and the Yelp API storing 5,000+ restaurant records in DynamoDB with Elasticsearch querying, delivering personalized recommendations via email within 2 minutes.
AWS LambdaAmazon LexDynamoDBSQSYelp API
5,000+
Restaurants
<2 min
Email delivery
Full Stack
CollabDesk
All-in-One Collaboration Platform
Unified multi-workspace platform for academic and professional teams featuring Kanban boards, calendars, resource repos, and AI-powered discussion summarization to eliminate tool fragmentation.
ReactDjango RESTSupabasePostgreSQLGemini AITravis CI
AI
Auto-summaries
4+
Native tools
ML & AI
Smart Traffic Light Controller
Deep Reinforcement Learning Agent
Adaptive traffic signal controller trained with Deep Q-Learning in SUMO simulation, reducing average intersection queue length by ~80% and cutting average delay from 20s to ~8s.
PyTorchDeep Q-LearningSUMOTraCI API
~80%
Queue reduction
1,000+
Vehicles simulated
ML & AI
CIFAR-10 Image Classification
Custom ResNet from Scratch
Designed and trained a 4.7M parameter ResNet (4-4-3 residual config) from scratch, benchmarking optimizers and schedulers with an aggressive augmentation pipeline.
PyTorchResNetCNNsSGDMultiStepLR
93.1%
Val accuracy
4.7M
Parameters
ML & AI
LoRA Fine-Tuning of RoBERTa
Parameter-Efficient NLP Fine-Tuning
Fine-tuned RoBERTa-base on AG News using LoRA adapters, training only 0.7% of parameters. Combined contextual data augmentation with student–teacher knowledge distillation.
HuggingFacePEFT / LoRARoBERTaDistillation
84.8%
Test accuracy
0.7%
Params trained
Computer Vision
Multi-Organ Nuclei Segmentation
Instance & Semantic Segmentation
Used Mask2Former (Swin-Base backbone, pre-trained on COCO panoptic) for joint instance and semantic segmentation of cell images, with augmentation to address severe class imbalance.
Mask2FormerPyTorchSwin TransformerPanoptic Seg
Panoptic
Segmentation
COCO
Pre-trained
Computer Vision
GeoGuessr
Visual Geolocation from Street Images
Predicted US state and GPS coordinates from 4 street-view images using StreetCLIP embeddings, with a custom State Classifier and GPS Regressor for fine-grained geolocation.
StreetCLIPPyTorchGPS RegressionCLIP Embeddings
96.6%
GPS score
90.5%
State accuracy
Cloud & DevOps
Intel® Trust Authority
SaaS Remote Attestation Platform
Industry-first independent Trust Authority for verifying compute asset trustworthiness. Provisioned secure cloud environments and automated CI/CD pipelines across AWS and Azure.
TerraformAnsibleHelmGitHub ActionsAWS / Azure
99.95%
Uptime
3 min
Tenant onboarding
IoT & Edge
Intel® Edge Insights for Industrial
Industrial IoT Microservices Platform
Pre-validated, ready-to-deploy software reference design for smart, scalable factory solutions — providing the infrastructure layer for video and time-series data ingestion via microservices and edge computing.
DockerKubernetesPythonPytestSnyk
96%
Test coverage
6+
Microservices
Blockchain
BlockTalk
Decentralized Chat Application
Web3-powered chat platform on Ethereum with IPFS storage supporting 1-on-1 and group messaging with censorship-resistant, wallet-authenticated access.
SolidityReact.jsEthers.jsIPFSMetaMask
100%
Decentralized
0
Central servers

Publications

IEEE · 2022
Smart Traffic Light Controller using Deep Reinforcement Learning
3rd International Conference for Emerging Technology (INCET) · Belgaum, India
A. Abhishek, P. Nayak, K. P. Hegde, A. Lakshmi Prasad, K. S. Nagegowda — doi: 10.1109/INCET54531.2022.9824501

Proposed a deep reinforcement learning approach for adaptive traffic signal control, trained using DQN in a SUMO simulation environment. Demonstrated significant reductions in queue length and average intersection delay compared to traditional static timing strategies.

IEEE Xplore

Awards & Recognition

Graduate School of Engineering Scholarship
Merit scholarship upon admission to NYU, recognizing exceptional academic achievements and potential for excellence in engineering studies.
Intel Department Recognition Award · Y2023 Q2
For exemplary achievements towards delivering a distributed computing platform for the intelligent network and edge, as part of Intel® Edge Insights for Industrial.
Intel Flex Fearless Innovation Award · Y2022 Q3
For excellent work and role modelling on Innovation activities at Intel Flex.
Prof. MRD Merit Scholarship Recipient
Ranked in the top 20% at PES University

Get In Touch

Let's build something interesting together.
Currently open to full-time SDE and ML Engineering roles starting Summer 2026. If you have an interesting problem or just want to chat, my inbox is always open.
Email
Email
aa12037@nyu.edu
Phone
Phone
(516) 646-8996
LinkedIn
LinkedIn
abhishek-adinarayanappa
GitHub
GitHub
Abhi270600