SYSTEM.INIT
Aakash
Munawar
Computer Science student at Rutgers, focused on databases, systems, and applied AI. Building toward my first internship.
Open to internships
About
I'm a CS major at Rutgers working through the core of the discipline — data structures, algorithms, computer architecture, databases, and applied AI. Still figuring out exactly where I want to focus.
SQL & Database Design
Data Structures
Algorithms
Computer Architecture
Discrete Math
Python
Java
Projects
01
2026
A SQL analysis of nearly 480,000 real Chicago Police Department incident reports (Sept 2024 – Aug 2026), built from a normalized relational database designed from scratch — not a flat spreadsheet.
- Reported crime fell about 5% year over year, with a clear seasonal cycle — lowest in February, highest in late summer/early fall
- Incidents peak around 3 p.m. and are rarest around 5 a.m.
- Theft dominates downtown and the North Side (up to 41% of incidents in some districts); battery leads across the South and West Sides
- Arrest rates vary hugely by crime type — almost all narcotics reports lead to an arrest, versus under 1 in 10 thefts
- Domestic-related incidents range from 4% to over 30% of reports depending on the neighborhood
02
2026
A command-line pathfinder for Rutgers New Brunswick, built on a graph of 24 real campus locations from OpenStreetMap. Finds shortest routes using Dijkstra's algorithm implemented from scratch with a binary-heap priority queue — not a library.
- Models real Rutgers geography — every coordinate is real OpenStreetMap data, with no simulated or made-up values
- Dijkstra's algorithm and the priority queue are hand-built, not imported from a graph library
- Cross-river routes correctly go over the real Landing Lane Bridge instead of cutting through the Raritan River
- Verified for correctness — every possible route in the graph was checked against an independent Floyd-Warshall algorithm, and all tests pass
- Also answers "what's nearest?" queries, like the closest dining hall or library from any location, ranked by route distance
03
2026
A command-line tool that finds every conflict-free weekly schedule from a list of real Rutgers courses using a backtracking algorithm written from scratch, then ranks results by preferences like fewest campus switches or latest start time.
- Uses real, current Fall 2026 Rutgers course data pulled from the university's public Schedule of Classes API
- Backtracking implemented from scratch — prunes conflicting branches immediately instead of checking every combination like brute force
- Measured against brute force directly: up to 5.9x faster and skips generating 87% of combinations on a 5-course search
- Verified for correctness — 18 automated tests confirm backtracking and brute force find the identical set of schedules on real course data
- Ranks results by real, useful criteria: fewest days on campus, fewest gap hours, latest start time, or fewest switches between campuses
04
2026
A full-stack Java web app where Rutgers students add their courses and weekly free time, then get ranked recommendations for study groups to join, or classmates to start one with, from a matching algorithm written from scratch. Demo login: recruiter@example.com / demo1234 (or use the "Try the demo" button).
- Matching algorithm hand-built in Java: free time is stored as sorted, merged intervals and intersected with a linear two-pointer merge; a group's time is when every member is free, and only blocks of an hour or more count
- Checked against an independent minute-by-minute brute-force model: 3,000 random schedule pairs and 1,575 ranked recommendations matched exactly, and all 41 JUnit tests pass
- Measured at 1.3 ms median per recommendation run for 5,000 simulated students (about 1,500 candidates each), for the algorithm alone; time grows linearly with candidates
- Spring Security with stateless JWT sign-in and BCrypt passwords; group rules enforced on the server, with a row lock so two students can't take a group's last seat
- Uses the real Fall 2026 Rutgers CS and Math course catalog (148 courses); the demo's sample classmates are labeled "(demo)" and kept in a separate sandbox that real students never see