Skills & Technologies
41 skillsWork Experience
Software Engineer Intern
IPOTHESIS RESEARCH PRIVATE LIMITED
- Engineered an end-to-end ML trading signal pipeline processing 50,000+ daily market data points, engineering 25+ technical indicators, training ensemble models, and generating signals with 65% directional accuracy.
- Delivered robust backtesting frameworks evaluating 15+ trading strategies across 3 years of historical data; automated A/B experiments comparing 8 model variants, improving Sharpe ratio by 0.4.
- Standardized dataset splits (70/15/15 train/val/test), 20+ feature configurations, and versioned training artifacts, reducing experiment setup time by 60% and improving reproducibility.
- Shipped and maintained 3 production MERN applications serving 500+ daily users, building 12+ RESTful APIs with Node.js and Express, delivering sub-200ms response times on 90% of endpoints.
- Collaborated with 4-person cross-functional team to ship 20+ production features over 2 sprints, achieving 99.5% system uptime and improving frontend load time by 35%.
Projects
Real-Time UPI Fraud Detection System
- Architected and deployed a real-time fraud detection pipeline processing 10,000+ transactions/hour with 100ms latency, ingesting streaming events via Kafka across 3 consumer groups for scalable, fault-tolerant detection with 99.9% uptime.
- Constructed a Redis-backed behavioral state engine tracking 6 fraud signals (velocity, amount, unique beneficiaries, fan-in patterns) using 1M+ TTL-based keys across 5-minute sliding windows, reducing false positives by 30%.
- Established an explainable rule-based fraud engine (8 heuristic rules) combined with Logistic Regression (AUC 0.87) using weak supervision on 100K+ synthetic transactions to produce hybrid risk scores.
- Orchestrated offline ML training on 250K+ labeled transactions and online inference pipelines with 15+ consistent features, achieving 82% precision and 78% recall in production scoring.
The Operating System Project
- Constructed a 32-bit i686 OS from 8,000+ lines of C and assembly with a two-stage bootloader (512 bytes stage-1) that detects memory via E820 BIOS calls and passes boot parameters into a 64KB kernel image.
- Integrated a FAT12/16/32 filesystem reader supporting 4KB cluster sizes in the stage-2 loader, along with an ELF loader parsing 10+ section headers to locate and jump to /boot/kernel.elf.
- Configured core CPU setup including GDT (5 segments), IDT (256 interrupt vectors), 32 ISR handlers, 16 IRQ handlers with an 8259 PIC driver, and x87 FPU initialization achieving 2ms boot time.
- Created a VGA-based GUI (320x200 16-color mode) and input stack with PS/2 keyboard (101-key layout), PS/2 mouse driver, and 3 desktop-style applications (text editor, file browser, calculator).
AI Technical Interviewer
- Architected a full-stack technical interview platform serving 100+ test sessions with adaptive difficulty scaling across 5 levels using Next.js frontend and FastAPI backend (8 REST endpoints) with 99% uptime.
- Deployed a 4-stage LangChain pipeline using Google Gemini for dynamic question generation (3 questions/round) and answer evaluation, processing 500+ candidate responses with 92% evaluator agreement vs human reviewers.
- Crafted an intelligent difficulty adaptation algorithm analyzing 7 scoring dimensions with normalized scores (0-100 scale) from previous answers, improving candidate engagement by 45% through personalized progression.
Leaderboard Standings
Leaderboard Position Pending
Global test scores, peer standing percentiles, and algorithm leaderboard ranks are updated dynamically.
Assessment Highlights
Assessments Not Completed
Coding evaluations, system assessment results, and conceptual score badges will appear here after taking a test.
AI Collaboration Score
AI Collaboration Score Pending
Developer coding behavior, assistant cooperation, and AI pair-programming indicators are evaluated during live coding sessions.
Role Compatibility Profile
Role Compatibility Analysis Pending
Custom matching reports, candidate role compatibility percentiles, and core engineer strength profiles are processed once conceptual code screenings are complete.
Achievements
Real-Time UPI Fraud Detection System
Architected a real-time fraud detection pipeline processing 10,000+ transactions/hour with 100ms latency.
About Details
Professional Bio
Software Engineering student with expertise in ML pipelines, full-stack development, and systems programming. Experienced in building scalable trading systems, real-time fraud detection, and AI-driven platforms. Passionate about high-performance computing and robust system architecture.
Bachelor of Technology in Computer Science in Computer Science
Bennett University (2022 - 2026)