Scott Alessio, Computer Engineer

Data Center Operations Engineer at Bloomberg L.P.

Jan 2026 - Aug 2026 · New York, NY & Northern New Jersey

Supported the physical and logical lifecycle of a global, high-availability data center fleet of 40,000+ multi-vendor x86 servers, enterprise routers, and switches across paired production sites.

• Investigated infrastructure and hardware faults in Splunk, building searches across system and network logs to isolate root cause and surface recurring failure patterns across the fleet.
• Performed hardware and OS-level troubleshooting across a multi-vendor server fleet, isolating faults at the component, firmware, and network layer to restore availability within service-level targets.
• Troubleshot pod-to-pod connectivity failures across Kubernetes clusters, tracing east-west traffic paths and inspecting Calico network policy to separate policy denials from node and upstream network faults.
• Executed server and network device installations, migrations, and secure decommissioning under formal change management, maintaining access controls and authoring the runbooks that standardized recurring work.
• Partnered with network, systems, and facilities teams to sustain redundancy and high-availability targets across the fleet.

Technologies: Splunk, Kubernetes, Calico, Linux, x86 Server Hardware, Out-of-Band Management, Change Management, Enterprise Routing & Switching

Information Technology Support Assistant at Ridgewood Country Club

May 2022 - Dec 2025 · Paramus, NJ

Delivered end-to-end IT support for staff endpoints, network infrastructure, and connected systems across a multi-building campus.

• Deployed and hardened an OpenVPN server on a cloud-hosted Linux environment, enabling encrypted remote access for staff and eliminating unsecured remote connection methods.
• Administered user accounts, endpoint configuration, and Microsoft 365 services, handling onboarding, offboarding, and access provisioning.
• Diagnosed network, wireless, and IoT faults across the campus and applied operating system and firmware patches, sustaining uptime while reducing exposure to known vulnerabilities.
• Delivered comprehensive IT support for host devices, keeping staff operations running across a multi-building campus.

Technologies: OpenVPN, Linux, Microsoft 365, Windows, 802.11 Wireless, IoT, Patch Management, Endpoint Hardening

Network & AI Research Lead at New Jersey Institute of Technology

May 2024 - Jun 2025 · Newark, NJ

Led an 8-person research team building LLM-driven automation for Cisco network device operations, from prompt design through hardware integration and evaluation.

• Co-authored and published 2 peer-reviewed papers on prompt engineering efficacy for automated network configuration and troubleshooting.
• Engineered a Linux-based, hardware-integrated LLM assistant that automated device management tasks, cutting emergency staffing costs 12% in internal testing.
• Directed evaluation methodology, code review, and publication timelines across the team, analyzing results to raise accuracy and reduce misconfiguration risk.
• Curated and analyzed key findings to optimize LLM prompts for network configuration tasks.

Technologies: Python, Linux, Cisco IOS, Large Language Models, Prompt Engineering, Network Automation

Enterprise Network Security Design

A segmented enterprise network simulated end to end in Cisco Packet Tracer, hardened against Layer 2 attacks.

Designed and simulated a segmented enterprise network of routers, switches, and hosts in Cisco Packet Tracer, modelling the access, distribution, and core layers of a realistic corporate topology.

Implemented VLAN subnetting to isolate broadcast domains and applied port security on the access layer to mitigate MAC spoofing and CAM-table flooding, the two Layer 2 attacks that most reliably break a flat network.

The exercise was as much about failure modes as design: each control was validated by attempting the attack it was meant to stop and confirming the switch behaved as configured.

Technologies: Cisco Packet Tracer, Cisco IOS, VLANs, Subnetting, Port Security, Routing & Switching

AWS Image Recognition Pipeline

Two EC2 instances running in parallel, chained through SQS, performing object and text recognition on an S3 image set.

A distributed image-recognition pipeline built on AWS with two EC2 instances working in parallel, written in Java against the AWS SDK v2.

Instance A reads images from a public S3 bucket, runs object detection through Amazon Rekognition, and pushes the index of every image containing a car (above 80% confidence) onto an SQS queue, then pushes a sentinel value to close the stream. Instance B consumes indexes as they appear, downloads each image, runs text detection, and writes a report of every image containing both a car and readable text.

The interesting constraints were ordering and start-order independence: a FIFO queue guarantees the sentinel arrives after every real index, and both applications create the queue idempotently on startup so either instance can be launched first.

Technologies: AWS EC2, Amazon S3, Amazon SQS (FIFO), Amazon Rekognition, AWS IAM, Java, Maven, Linux

Secure Serial Data Transmission System

A serial communication system in C with integrity checking, running on a co-designed RISC-V circuit network.

Built and tested a serial communication system in C with integrity checking to detect and prevent data corruption in transit.

Co-designed the supporting circuit network using RISC-V architecture, integrated circuits, and discrete transistors. The software and the hardware it ran on were developed together rather than one against a fixed spec.

Presented the finished system to a formal technical review panel.

Technologies: C, RISC-V, Serial Communication, Digital Logic, Integrated Circuits, Hardware Debugging

Clustering & Outlier Detection From Scratch

K-means, hierarchical clustering across four linkage methods, k-NN outlier detection and Silhouette scoring, implemented in pure Python.

A data-mining study implementing the core unsupervised algorithms from first principles in pure standard-library Python, with no numerical libraries in the implementation path.

Covered k-nearest-neighbour outlier detection, K-means, agglomerative hierarchical clustering across four linkage strategies, and Silhouette coefficient evaluation, run against three synthetic three-dimensional datasets.

Results were independently cross-checked against established library implementations to confirm the from-scratch versions produced equivalent clusterings.

Technologies: Python, K-Means, Hierarchical Clustering, k-NN, Silhouette Analysis, Data Mining

Solar Flare Prediction: ML Reproduction Study

Reproduced ten published solar machine-learning tools on modern hardware and documented which ones still build.

A reproducibility study covering ten published solar machine-learning tools, rebuilt and re-run from their original repositories against a single shared Python 3.10 and TensorFlow 2.10 environment.

The result was as much negative as positive: several tools pinned to TensorFlow 1.x and Python 3.6 are effectively unbuildable on current systems, while the rest could be unified under one environment with a small number of compatibility patches.

Paired with a close reading of an LSTM-based flare-prediction paper and presented as a full technical deck.

Technologies: Python, TensorFlow, Keras, scikit-learn, LSTM, Reproducibility

Partitioning Prompts for Higher Efficacy in Network Design with Large Language Model

Proposes breaking a dense network-configuration prompt into a sequence of smaller, focused tasks, and shows that this partitioned approach yields more accurate and consistent LLM responses than asking for everything at once.

Optimizing LLM Prompts for Automation of Network Management: A User's Perspective

Introduces a scoring policy for grading LLM-generated network designs and router configurations, then measures how prompt specificity, and schematic rather than text prompting, affects error rates across three levels of network complexity.

Master of Science, Cyber Security and Privacy, New Jersey Institute of Technology

May 2025 - Aug 2026 · GPA 3.85 / 4.0

Coursework: Security & Privacy in Computer Systems, Internet & Higher-Layer Protocols, Cloud Computing, Data Mining, Network Security

Bachelor of Science, Computer Engineering, New Jersey Institute of Technology

Sep 2021 - May 2025 · GPA 3.62 / 4.0

Coursework: Computer Networks, Cybersecurity, Operating Systems, Computer Architecture & Design, Digital Data Communications, Microprocessor Systems Design

High School Diploma, Paramus High School

Graduated 2021 · GPA 4.0 / 4.0