kanishk wagh
Open to work0 years experienceAI Engineer
AI Engineer | LLM Systems | Agent Loops | Harness Engineering | Agentic Workflows | AI Security
Standing Rank
Rank Not Available
Developer Badges
No badges earned yet
Skills & Technologies
35 skillsWork Experience
AI & Automation Engineer Intern
Crown Solution
- Designed and deployed autonomous AI pipelines using Python, FastAPI, PostgreSQL, and Redis — eliminating 10–15 hours/week of manual SEO engineering work per team member ( 80% effort reduction ).
- Implemented agentic workflow patterns with multi-step autonomous task execution: crawling → issue classification → fix recommendation, operating with zero human intervention.
- Engineered closed agent loops with observe-classify-act cycles, enabling autonomous multi-step SEO task execution and self-correcting pipeline behaviour without human triggers.
- Built evaluation harnesses to test AI pipeline outputs against expected SEO metrics, catching regressions before production deployment and ensuring pipeline reliability at scale.
- Integrated LLMs and external APIs for intelligent content analysis, keyword classification, and AI-driven optimization workflow orchestration across the full data pipeline.
- Applied secure API integration practices: token management, rate limiting, and input validation to ensure AI pipeline integrity and data security.
Projects
Lumiqe — AI Color Analysis & Personal Styling Platform
- Built a 9-step computer vision AI pipeline (BiSeNet, MediaPipe, K-Means++, Delta-E CIE2000) for lighting-invariant skin-tone classification across 12 colour seasons in under 2 seconds — production-deployed on AWS EC2.
- Integrated an AI stylist chat powered by Llama 3.3 70B (Groq) with server-side JWT proxy architecture ensuring auth tokens are never exposed client-side — secure LLM integration design pattern.
- Implemented Redis rate limiting, input sanitisation, and output filtering on AI endpoints — applying AI security guardrails to prevent prompt injection and content policy violations in production.
- Shipped full SaaS product: Stripe billing, wardrobe management, price-drop alerts, referral system, and EN/ES localisation.
Competitor Analyzer Agent — Autonomous Market Intelligence System
- Built an autonomous AI agent monitoring competitor data in near real-time; integrated SHA-256 content diffing for tamper-aware change detection and LLM-based classification for structured insight generation.
- Designed agentic pipeline with input validation and output sanitisation to prevent data injection attacks and ensure structured, safe AI outputs.
- Built a lightweight evaluation harness to validate agent output quality — automatically flagging classification drift and pipeline degradation before surfacing insights to downstream consumers.
TrackML — Particle Track Reconstruction
- Built a Transformer-based GNN for 3D particle trajectory reconstruction on high-dimensional physics data with no predefined solution — resolved 1000:1 class imbalance through targeted data architecture decisions.
- Optimised training pipeline with FP16 mixed precision; evaluated using precision, recall, and F1-score.
Drone Detection System — Real-Time Computer Vision
- Developed a computer vision detection pipeline on a 2,783-image custom dataset achieving mAP@0.5 of 0.89 at 10–12 ms/frame — suitable for live deployment in security and surveillance contexts.
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
HackVision Hackathon
2nd Runner-Up, HackVision Hackathon (Top 1% of 200+ teams).
Jio Institute x GDG Cloud Mumbai Hackathon
Thematic Challenge Winner, Jio Institute x GDG Cloud Mumbai Hackathon.
About Details
Professional Bio
AI Engineer specialising in LLM systems, autonomous agentic workflows, and AI security. Experienced in building production-grade AI systems end-to-end, including closed agent loops, evaluation harnesses, and secure AI deployment. Strong foundation in responsible AI and OWASP LLM security principles.
B.Sc. in Data Science in Data Science
Thakur Shyamnarayan Degree College (2023 - 2026)