Omscs

OMSCS: Game AI and Software Analysis (Spring 2026)

Spring 2026 was a two-course semester. Coming into it, I knew it’d be a stretch — Game AI on the creative side, Software Analysis on the formal side — but I wanted to see how they’d play off each other while balancing a full-time QA role. Turned out they complemented each other better than I expected. CS7632 — Game AI Game AI is fundamentally different from academic AI. You’re not trying to prove optimality or solve an intractable problem in the general case — you’re making a player have a good time. That means cutting corners, embracing approximations, and yes, cheating when the player won’t notice. The entire course is built around this philosophy.

Saturday, May 2, 2026 | 6 minutes Read

OMSCS: CS 7650 Natural Language (Fall 2025)

Taking Natural Language Processing in Fall 2025 was a mixed experience. I’d taken a similar course during undergrad, and there was significant overlap — enough that a good portion of the material felt like revisiting ground I’d already covered. That said, the deeper dive into Transformers and attention mechanisms was new and genuinely useful. Course Snapshot Course CS 7650 Natural Language Term Fall 2025 Credits 3 Why I Took It I’ve always been drawn to language — how we communicate, how meaning is constructed, why the same sentence can mean different things in different contexts. NLP sits at the intersection of that human interest and the machine learning techniques I’ve been building up through OMSCS. Plus, with LLMs becoming ubiquitous, I wanted a proper academic foundation for understanding what’s happening under the hood — not just the API-calling layer.

Thursday, December 11, 2025 | 5 minutes Read

OMSCS: CS 6290 High Performance Computer Architecture (Spring 2025)

CS 6290 was one of those courses where I came out knowing exactly how a processor works under the hood in a way I hadn’t before. It covers the full stack of modern processor design — from the simplest pipelining all the way up to multi-core coherence protocols and memory consistency models. It’s dense, quantitative, and rewards meticulous attention to detail. What The Course Covered The course followed a logical arc that built up in complexity. It started with fundamental metrics — the Iron Law of Performance (execution time = instruction count × CPI × cycle time) and Amdahl’s Law, which became the recurring mental models for evaluating every design tradeoff.

Thursday, May 1, 2025 | 5 minutes Read

OMSCS: GIOS and High Performance Computing (Fall 2024)

Fall 2024 was my first semester in OMSCS, and I took on two courses: CS 6200 Graduate Introduction to Operating Systems (GIOS) and CSE 6220 Intro to High Performance Computing (IHPC). It was a heavy start, but the two courses complemented each other well — GIOS covered the abstractions and mechanisms of modern systems, while IHPC pushed into the algorithmic side of parallel computation. CS 6200 — Graduate Introduction to Operating Systems GIOS was the anchor course of the semester — a deep, practical dive into how operating systems work, from single-node mechanisms through to distributed systems. The course is structured around the lens of abstractions, mechanisms, and policies, which became a recurring mental model I still use when thinking about systems design.

Sunday, December 15, 2024 | 6 minutes Read

OMSCS: CS 6601 Artificial Intelligence (Spring 2024)

CS 6601 Artificial Intelligence was my first OMSCS course in Spring 2024. Starting the program with a single course felt like the right call, and AI turned out to be a solid choice — it’s a broad survey that covers search, probability, decision theory, and machine learning, all tied together by the thread of building intelligent agents. What The Course Covered The course is organised around the idea of an intelligent agent perceiving its environment and acting upon it. It started with search — the foundation of many AI problems. Uninformed search (BFS, DFS, uniform-cost) then informed search with A* and the critical concept of admissible and consistent heuristics. The heuristic quality directly determines whether A* finds the optimal path efficiently or explores the entire space. Local search methods like hill climbing and simulated annealing followed, plus constraint satisfaction problems with backtracking and arc consistency.

Thursday, May 2, 2024 | 4 minutes Read