Education
USC Viterbi B.S./M.S. (Computer Science, Spring 2029)
Forbes-Featured Engineer · USC Viterbi
Building resilient software systems that survive real traffic, real failure, and real scale.
I'm Mo Shirmohammadi, a software engineer focused on distributed systems, data infrastructure, and applied ML. I design production-minded systems that prioritize reliability, observability, and measurable user impact.
Scale
AI student-support chatbot used by 30,000+ students
Impact
640+ daily queries, 50% lower latency, 99.9% uptime
Recognition
Forbes-featured · 1st Place HTCC 2024
Flagship Work
Eight projects that prove systems depth, product execution, and technical range.
DataSystemsML
Surgical Data Mesh Platform
Production-style data platform for robotic surgery telemetry with durable ingest, real-time safety alerting, and a lakehouse analytics path.
- Replay-safe ingest with event-level idempotency and quality gates
- Streaming + lakehouse path (Kafka -> Spark -> Iceberg)
- Serving stack tuned for timeline, search, and flexible case documents
Tech Stack
PythonJavaKafkaSparkIcebergAirflowdbtKubernetesTerraform
Impact Signals
- Data PatternReal-time + batch fusion
- Storage ModesOperational + analytical
- DeploymentCloud-native multi-service
DataMLSystems
SurgGraph Pipeline
Graph-enhanced robotic surgery analytics pipeline prototype focused on bottleneck detection, recovery-risk modeling, and workflow intelligence.
- Streaming telemetry workflow with transformation and dashboard serving
- Procedure-graph path analysis for bottleneck and route optimization
- Risk-model pipeline with fairness disparity checks
Tech Stack
PythonKafkaAirflowdbtDuckDBDockerAWS
Impact Signals
- Core FocusGraph-aware analytics
- Data FlowStream-to-dashboard
- Operating ModePrototype + experiment loop
Systems
Distributed Task Queue
Production-grade distributed queue in pure Python with WAL persistence, retries, deduplication, circuit breaker protections, and DAG scheduling.
- Priority scheduler with replay-safe checkpoints and crash recovery
- Retry/backoff + dead-letter queue + dedup safeguards
- DAG and cron execution with P50/P95/P99 latency instrumentation
Tech Stack
PythonWALConcurrencyReliability
Impact Signals
- ReliabilityCrash recovery + replay
- SchedulingPriority + DAG + cron
- GuardrailsCircuit breaker + DLQ
SystemsData
Database Replication Engine
PostgreSQL-inspired replication system from scratch with WAL propagation, leader election, automatic failover, and split-brain handling.
- Leader-follower replication over ordered WAL streams
- Failover workflow with election and quorum-style recovery
- Sync/async/semi-sync modes to surface consistency tradeoffs
Tech Stack
TypeScriptReplicationConsensusFault Tolerance
Impact Signals
- Replication ModesSync / Async / Semi-sync
- Failure ModelLeader loss + recovery
- DurabilityLog-driven state rebuild
SystemsData
KV Store Engine
Java key-value engine with a custom open-addressing hash table, TTL eviction strategies, and atomic snapshot persistence.
- Custom hash core with collision-aware probing and fast lookup
- TTL eviction via lazy + active cleanup strategies
- Atomic snapshot save/restore for consistent recovery
Tech Stack
JavaSpring BootData StructuresStorage Engine
Impact Signals
- Storage CoreNo built-in map
- EvictionTTL + active sweep
- RecoverySnapshot restore
Systems
Load Balancer from Scratch
Java load balancer implementing six routing algorithms with health checks, concurrency controls, and failure simulation.
- Algorithm suite including round-robin, weighted, and least-connections
- Health probes and fail-fast routing for unavailable backends
- Concurrent request handling with throughput/latency measurements
Tech Stack
JavaNetworkingDistributed SystemsConcurrency
Impact Signals
- Algorithms6 routing strategies
- OperationsHealth-check aware
- PerformanceLatency/throughput tracked
Systems
Raft Consensus Simulator
Consensus simulator featuring deterministic leader election, log replication behavior, and fault injection scenarios.
- Term-based election model with vote coordination
- Heartbeat and replication mechanics under partial failures
- Deterministic simulation for repeatable debugging and learning
Tech Stack
Distributed SystemsConsensusFault Injection
Impact Signals
- Core BehaviorElection + heartbeat
- Testing ModeDeterministic scenarios
- GoalConsensus intuition
MLSystems
Transformer LM from Scratch
GPT-style decoder model implemented from first principles with manual backpropagation and comprehensive test coverage.
- Decoder-only transformer stack with explicit attention flow
- Manual gradient pipeline for full training transparency
- Comprehensive test suite validating numerical and functional correctness
Tech Stack
PythonDeep LearningBackpropagationTesting
Impact Signals
- ArchitectureDecoder-only transformer
- Learning DepthManual gradient path
- QualityFull test coverage mindset
Interactive Lab
Playable systems simulations that mirror real design tradeoffs.
Shows how priority + retry strategy shifts throughput, tail latency, and dead-letter pressure.
Controls: Priority weight · Failure rate · Backoff intensity · Tick simulation
- task-184mscomplete
- task-2105mscomplete
- task-684mscomplete
- task-5126mscomplete
- task-8105mscomplete
- task-3213msretrying
- task-4168mscomplete
- task-7147mscomplete
Complete
7
Retrying
1
Dead-letter
0
P95 Latency
168ms
Project Catalog
Additional systems and infrastructure builds.
SystemsData
LSM-Tree Storage Engine
WAL, memtables, SSTables, compaction, and Bloom filters.
SystemsData
Postgres-backed Pub/Sub Queue
Durable messaging with consumer-group offsets and ACID semantics.
SystemsData
SQL Database Engine
Parser, executor, indexing, persistence, and REPL.
Systems
HTTP Server from Scratch
Multithreaded HTTP/1.1 server over raw TCP sockets with middleware and routing.
Systems
Container Runtime from Scratch
Docker/OCI runtime simulator with namespace and cgroup mechanics.
DataSystems
Search Engine from Scratch
Inverted index with BM25 ranking, boolean queries, and lightweight UI.
Systems
DNS Resolver from Scratch
Iterative DNS resolver over raw UDP sockets from root servers.
Systems
Memory Allocator Simulator
Heap placement strategies: first-fit, best-fit, and worst-fit.
Systems
Myers Diff from Scratch
Myers diff algorithm implementation with unified diff output.
DataSystems
B-Tree Index from Scratch
B-Tree index implementation with CLI visualization.
SystemsFull-Stack
Raytracer from Scratch
C++17 ray tracer with shading, reflections, anti-aliasing, and multithreaded rendering.
MLSystemsData
Robot Log Analyzer
Automated anomaly detection and root-cause triage for robotics V&V logs.
Experience
Shipping at scale across campus tech, enterprise systems, and education.
Founding Software Engineer
Pasadena City College
Dec 2024 - Aug 2025
- Shipped a full-stack generative AI chatbot serving 30,000+ students and handling 640+ daily queries.
- Cut response latency by 50% with backend rate limiting, caching, and optimized OpenAI API orchestration.
- Improved answer accuracy from 38% to 66% through interaction-log-driven prompt iteration.
- Built CI/CD with GitHub Actions and 50+ automated tests, reducing deployment errors by 60%.
- Launched RAG with vector search over 900+ documents and reduced hallucinations by 40%.
- Sustained 99.9% uptime with centralized Sentry error telemetry and sub-hour incident response.
Verkada Systems Technician
Access Tech Security
Apr 2023 - Present
- Resolved hardware-software integration failures across 150+ client sites with a 90% first-visit resolution rate.
- Configured enterprise VLAN/subnet/static-IP network setups for reliable IoT cloud connectivity.
- Automated cross-platform data synchronization with Python against Sonar, ROE, and GraphQL APIs.
Lead Sensei of Curriculum Engagement
Code Ninjas
Jan 2023 - Feb 2025
- Designed a parent-teacher conference platform used by 80+ families.
- Led Python/C# game-development camps and curriculum delivery for youth learners.
Skills Matrix
Core toolkit for backend, data, and ML systems.
Languages
PythonJavaScriptTypeScriptJavaC++SwiftSQLBash
Backend & APIs
Node.jsExpressDjangoSpring BootFastAPIGraphQL
Data & Infra
PostgreSQLMongoDBCassandraElasticsearchRedisKafkaSparkIcebergAirflowdbtDockerKubernetesTerraformAWS
AI/ML
OpenAI APIRAG PipelinesVector EmbeddingsTensorFlowPyTorchPrompt Engineering
About
I care about systems that are fast, fault-tolerant, and explainable under pressure.
I'm currently in USC's accelerated B.S./M.S. Computer Science track. My engineering style is pragmatic: design for reliability first, instrument every critical path, and let measurable outcomes drive iteration. Across projects, I aim to bridge theoretical depth and production constraints.
Outside class and internships, I build from-scratch engines and simulators to sharpen systems intuition: replication, scheduling, storage internals, consensus, networking, and model pipelines. That practice translates directly into cleaner, faster delivery on real teams.
Contact
Looking for a software engineering intern who ships fast and thinks in systems?
I'm targeting big-tech software engineering internships. If you're hiring for backend, distributed systems, data platform, or infrastructure-adjacent roles, I'd be glad to connect.