A V Karthikeya

Computer Science Engineer (B.Tech 2026) with a strong foundation in software tool development, scripting, and database-backed systems. Experienced in building and maintaining production-grade Python/Django applications, authoring and debugging automated evaluation pipelines, and working across Git/GitHub, Docker, and Unix/Linux environments. Comfortable balancing multiple critical-path, time-sensitive work streams and coordinating within a team to apply consistent engineering methodologies.

Seeking an entry-level Software Engineer role focused on tool development, process automation, and data-driven systems.

AI Trainer & Agent Builder (Freelancer) at Outlier.ai (2026-01 – Present)

• Design task ideas by defining scope, constraints, and complexity; author structured prompts for AI agent evaluation.
• Run prompts to generate comparable trajectories across multiple models and extract them into standardized task formats.
• Assess trajectories for safety failures and rubric violations; create custom evaluation rubrics covering behavior, tool usage, reasoning, and output quality.
• Rate and rank model performance to determine final preference rankings, directly contributing to AI model improvement.

Benchmark Task Author (Freelancer) at Handshake AI (2026-01 – 2026-12)

• Authored a full benchmark task (Project Dynamo) simulating a JavaScript monorepo semantic-versioning release pipeline across 30 interdependent packages, for evaluating AI coding agents.
• Built the task's Docker-based test environment, an oracle solution script, and an automated verifier in Python, then validated correctness locally before submission to an automated review pipeline (rubric review, duplicate detection, adversarial cheat-pass testing, and QC).
• Reverse-engineered and iterated on a rules engine that infers versioning/dependency-rewrite policy from historical release data, debugging edge cases around dependency graph traversal, package exclusion rules, and cascading version updates.
• Managed the full Git/GitHub contribution workflow (forking, branching, pull requests, CI re-triggers) and diagnosed environment issues across Docker, WSL2, and PowerShell.

Student Intern at AshnaAI (2025-08 – 2025-09)

• Contributed to development and improvement of AI agents using Generative AI and prompt engineering techniques.
• Designed instruction sets and conversational workflows to improve task-oriented response quality and reasoning structure.
• Optimized AI agent interactions for higher response accuracy, workflow reliability, and intelligent response generation.

Software Development Engineer Intern at Bluestock Fintech (2025-05 – 2025-06)

• Designed and developed a production-grade IPO web application providing real-time IPO insights including price bands, issue size, CMP, listing price, and downloadable DRHP/RHP documents.
• Built robust backend RESTful APIs using Django and Django REST Framework; managed data storage and scalability with PostgreSQL.
• Developed a clean, responsive frontend using React.js and Tailwind CSS; integrated real-time data updates with Angular and Chart.js dashboards.
• Performed EDA on financial market data from NSE India using Python, Pandas, and NumPy; analyzed market trends, volatility patterns, and trading behaviour.

B.Tech in Computer Science and Engineering – Presidency University, Bangalore (2022-01 – 2026-12)