📍 Delhi, India | 🤖 AI Systems Architect & Research Explorer | 🚀 GSoC '24 Alum, MATS Stage 2 Selectee & AI Safety Advocate

20 y/o AI undergrad at Newton School of Technology (B.Tech in CS & AI, expected '27), fusing backend resilience with AI innovation to craft scalable, ethical systems that push boundaries in agentic engineering and interpretability. From sharding cloud storage for fault-tolerant infinity to dissecting models for safety insights, I turn research into production-grade tools that empower devs while guarding against risks like over-reliance or biases. My path weaves open-source impact, global forums, and hands-on builds—always probing nuances like distributed failures, ethical deployments, and real-world edges to deliver solutions that scale with integrity.

🍭 DevXAnonymous (org) – Collaborative forge for dev tools, AI agents, and ethical experiments, sparking community-driven breakthroughs in accessible tech.

These flagship builds tackle real dev frictions with AI smarts and robust engineering—think seamless integrations, resilient architectures, and user-centric vibes. Each explores technical depths (e.g., sharding algorithms), nuances (e.g., privacy-preserving syncs), implications (e.g., democratizing opportunities), and edges (e.g., quota overflows or vetting biases). Pinned for impact, they're live and evolving.

• 🗄️ DriveMerge (collab) - Astute aggregator sharding data across clouds for unlimited, resilient storage. Masters transfers, quota ballet, and outage recovery—redefining distributed reliability with seamless UX.
• 👥 PEOPLE - Vetted gig platform bridging devs to quality opportunities with fair bidding and integrity checks. Navigates equity nuances, reshaping freelance dynamics while dodging spam edges.
• 📚 AXIOM - Holistic dev ecosystem for skill honing, assessments, and career trajectories. Infuses AI for tailored paths, pondering bias mitigations and personalization implications.
• 🔮 Oracle - API key oracle unveiling hidden powers and perils with smart scans. Delves into security implications, handling expired or leaked key edges like a pro.
• 💬 StoreChat - Effortless LLM chat vault syncing sessions to repos for reference gold. Tackles persistence subtleties, boosting workflow retention without format fiascos.
• 🛡️ SkillSheild - AI watchdog reversing tool dependency by tracking and rebuilding skills. Born from automation insights, spans psych impacts and custom metric outliers for sharper devs.
• 🌈 colour-fun - Instant backdrop wizard for solids/gradients—custom vibes for design flows. Explores color science nuances and accessibility ramifications.
• 🧠 ClasspilotAi - Real-time AI mentor crushing classroom doubts with pinpoint answers. Weighs ethical reliance, thriving in low-bandwidth edges.
• 🗂️ savespace - Clever organizer purging storage bloat intelligently. Probes data ethics, with privacy implications and massive dataset edges.
• 🖥️ adityakammati-Portfolio - Slick backend for my showcase site (frontend brewing at adigix.vercel.app). Scales with TS, eyeing expansion horizons.
• 🔄 uniloop - JS powerhouse for automation loops in dynamic apps—versatile for concurrency dances and adaptive scenarios.
• 📝 DevNotes - Coder's quick-note haven with search superpowers and integrations. Handles ephemeral ideas, dodging capture bottlenecks.
• 🛠️ FailFast - Startup scanner exposing flop risks early. Draws from venture wisdom, covering market shift edges and risk vectors.

Deep dives into AI frontiers, focusing on alignment, interpretability, and applied ML—these unpack models for transparency, forecast futures, and ensure grounded ethics. Vectoring from tech (e.g., feature engineering), context (e.g., safety puzzles), nuances (e.g., dataset biases), and implications (e.g., real-world anomalies).

• 📊 floor-segmentation-svm - Precision SVM for indoor mapping (96.75% accuracy). Harnesses multi-scale feats; debates overfitting nuances and deployment scales.
• 🤖 MATS-Summer-2026-Preparation - Curated arsenal of experiments and resources for AI alignment quests. Builds on Stage 2 wins, targeting mechanistic revelations.
• 🔬 GPT-2-Transformer-Training - Hands-on transformer tuning journey. Probes training subtleties and scaling ramifications.
• 🌈 skin-tone-classification - Tone detector for fashion recs. Tackles inclusivity biases and ethical hues.

Terminal titans for dev efficiency—diagnostics, tweaks, and snapshots that cut through complexity. Each packs power in simplicity, exploring CLI ergonomics, script robustness, and edge-handling like distributed chaos.

• 🔧 NetworkOptimizer - Suite for VPN polishes and speed surges. Fixes insecure setups, diving into network topologies and performance peaks.
• 🗂️ why-this-broke - Bug hunter capturing runtime states in distributed realms. Essential for debugging depths, handling concurrency quirks and scale stresses.

• Crafting AI agents & tools: Actively prototyping autonomous systems and dev boosters, iterating at vibe-speed.
• Hunting opportunities: Applying to top AI research gigs, fellowships, and roles to level up impact.
• Engaging dev ecosystems: Active on platforms like GitHub, LinkedIn, and forums—contributing, collaborating, and sparking discussions.
• Bridging research to builds: Turning alignment insights into practical agents for ethical, scalable AI.
• Global networking: Leveraging HPAIR connections for cross-cultural AI dialogues.

Coming soon—diving into AI agents, safety workflows, and dev hacks on adigix.vercel.app.

• GSoC '24 Alum (Creative Commons): Boosted open-source ML infra, mastering scalable dev.
• MATS Stage 2 Cleared: Advanced in ML Alignment & Theory Scholars, honing interpretability chops.
• HPAIR Delegate 2024/2025/2026: Harvard-picked for Asian & International Relations—forging global AI ethics ties.
• Completed Software Engineering at Disprz: Engineered uptime wins, slashing downtimes in edtech scales.
• 3rd Place, Stealth Fire Hackathon '24: For trailblazing AI prototype.
• GitHub Pro & Pull Shark x2: Merged PRs fueling ecosystems.
• 20+ Repos: AI/ML/dev arsenal with 1,415+ commits last year.

• HPAIR Selection Post - Buzz on 2024 delegate vibe, eyeing future waves.
• Disprz Journey Wrap - Reflections on internship triumphs, now complete.

"Innovate with guardrails—build agents that amplify humans ethically." From GSoC open ethos to MATS safety zeal, I engineer for resonance: systems that fail gracefully, AI that aligns deeply, tools bridging theory to deployment. Always multi-angling: tech feasibility, user nuances, ethical ripples, and societal edges.