I work on backend infrastructure, distributed systems, and AI/ML applications. I interned twice at
Google, on Cloud Infrastructure and on YouTube Ads. Before that I built enterprise tooling at Verizon
and studied quantitative problem solving at Jane Street. I graduated from UMass Amherst with a BS in
Computer Science in December 2025.
Software Engineering Intern · Cloud Infrastructure · Cambridge, MA
Worked on Google's global network infrastructure team, building real-time fault detection and
root-cause inference systems in C++. Focused on network telemetry, large-scale observability, and
production SRE tooling across Spanner and F1.
Built internal dashboards and data pipelines for YouTube's ads creative optimization workflows.
Worked across the full stack on Google's internal frameworks, integrated LLM-driven APIs, and tuned
backend query performance at scale.
Verizon
–
Software Engineering Intern · Mobile Infrastructure · Basking Ridge, NJ
Revamped Verizon's mobile device provisioning tool and optimized its backend database for millions
of daily device activations.
Jane Street
–
UNBOXED Fellow · New York, NY
Selected as 1 of 37 students internationally for the UNBOXED program. Studied market structure and
quantitative problem solving, and programmed in OCaml. Built an ETF trading simulator and a
real-time order book.
Projects
Agent Engine for Camera Effects
– Present
Multi-agent framework that turns natural language prompts into rendered visual effects using LLM
tool calling, WebGL, and MediaPipe. Includes an observability layer that tracks agent runs, token
latency, and tool-calling performance across thousands of prompts.
TypeScript
Go
React
gRPC
Cleverbot Chatbot
Transformer-based chatbot on Discord, running across multiple shards and serving 13,500+ servers and
54,000+ users. Scaled with async inference, Dockerized APIs, Kafka-based logging, and load-balanced
request routing.
Python
TensorFlow
Flask
Docker
Kafka
API Client Libraries (PyPI)
Open-source async API wrappers, async-cse and async-cleverbot, with 750,000+ combined downloads on
PyPI. Built with rate limiting, fault-tolerant retries, and async request handling.
Custom 3D multiplayer engine with multithreaded game logic, UDP networking over ENet, and spatial
partitioning that keeps the server on a lock-free tick under 5 ms.
C++
ENet
Spatial Partitioning
Bazel
Education
University of Massachusetts Amherst
BS, Computer Science · Manning College of Information and Computer Sciences
Former Committee Chair, Minutemen Alternative Investment Fund.
Skills
C++
Python
Go
Java
TypeScript
OCaml
gRPC / Protocol Buffers
SQL / Spanner
Docker / Kubernetes
GCP / AWS
Bazel / Terraform
Linux
Languages
English Native
Albanian Native
French Professional
Italian Professional
Contact
Open to roles in backend engineering, distributed systems, AI/ML infrastructure, and cloud platforms.