Rithish Reddy@TheLoneSynapse
5
Events attended
3
Submissions made
India
1 year of experience
About me
Hi, I’m Rithish Reddy An aspiring Information Security Specialist and first-year Cybersecurity student at Anurag University. I am driven by a passion to safeguard digital ecosystems and tackle modern tech challenges. Currently focused on mastering network defense protocols, threat analysis, and ethical security methodologies, I am actively building a strong technical foundation in systems architecture and programming. I’m eager to participate in CTF challenges, collaborate on innovative security projects
🤝 Top Collaborators
🤓 Latest Submissions
Fulcrum Service Recovery
Big delivery platforms run automated customer-recovery engines in-house. A 3–30 location restaurant group handles complaints by hand, from memory, under time pressure: slow, inconsistent, and refunds drift from policy. Fulcrum Service Recovery gives small groups that capability without handing decisions to a bot. Staff paste one or many complaints, optionally attach an anonymized order CSV, and pick a versioned written policy. AMD Developer Cloud Token Factory runs DeepSeek-V4-Flash (OpenAI-compatible API, temperature 0.2, JSON-only) to extract the issue, order refs and missing facts, and to draft a reply. Deterministic code then decides what is allowed: order verification, policy caps, never-automate rules, complaint-vs-record conflicts, and prompt-injection detection. Each case becomes a manager review packet: the draft (or why none was released), evidence refs for every claim, missing facts, inference provenance (backend, model, request ref) and an audit trail. Model error, timeout or bad JSON fails closed to manager review, with no silent fallback. Managers approve, reject or request info; decisions only record an audit event. Results on the live AMD backend (20-case synthetic eval): 20/20 correct routing, 100% packet completeness, 0 unauthorized side effects; median latency 6.24 s, p95 11.50 s on the shared endpoint. Built for shared capacity: 429 backoff honoring Retry-After, request timeouts, and a dedicated MI300X provider slot ready to configure. Next: MI300X deployment, a larger eval set, an owner policy editor, and a pilot with a restaurant group on its own written policy. All data is synthetic; no refunds, messages or customer contact. Stack: FastAPI, Pydantic, SQLite, Docker.
18 Oct 2026
Memory-Lane
Memory Lane is a gentle photo and memory companion built for older adults. It keeps the people, photographs and stories of a life in one calm place: a memory of the day on the home screen, a library of moments worth keeping, a circle of familiar faces, and stories to hear again. At its heart is a voice companion running on the AssemblyAI Voice Agent API. It answers from the person's own memories rather than the open web, and every action it takes is a small endpoint on the app itself — so when it saves a dictated note, the screen refreshes before the call has even ended. It can look someone up and open their card on screen, add a reminder to today, cancel or move a plan on any upcoming day, correct the year or name of a memory already saved, define a word, and search the web when a fact from outside the app is needed. Nothing is ever thrown away: taking a plan off a day hides it and writes a reviewable change log a family member can read, with who asked, when, and why. The whole app runs locally with zero configuration — no keys, no login — and setting two environment variables turns the same server into a multi-user app backed by Supabase auth and private photo storage.
30 Sep 2026
Bug Whisperer
🐛 Bug Whisperer — Find Mistakes in Your Code, Automatically Bug Whisperer is a custom skill built for IBM Bob during the IBM Bob 2.0 Hackathon (lablab.ai). It reads your Python code and tells you exactly what went wrong, where it went wrong, and how to fix it — in simple, friendly language, instead of a wall of raw traceback. Two Modes : Mode 1 — Fix Failing Tests 🔴→🟢 Triggers the moment your tests go red. Analyses every failing test in parallel, finds the root cause with file:line evidence, applies the smallest possible fix, and re-runs the full suite. If the fix doesn't work, it's rolled back automatically. Mode 2 — Review Your Code 🔍 Runs on fresh code before it ever breaks. Six parallel agents check for logic errors, security issues, missing error handling, concurrency bugs, resource leaks, and test gaps — returning a ranked report backed by evidence. The Mistake Box : Every finding is rendered as a clean terminal box showing WHERE the mistake is, WHAT went wrong, and step-by-step SUGGESTIONS to fix it. It plugs directly into VS Code via a task + problem-matcher, so the exact broken line gets a red squiggle with no extension required. Voice Mode 🎙️: Say "find the mistake in demo math" out loud and Bug Whisperer records your voice, transcribes it with IBM Watson Speech-to-Text, resolves the right file, runs it, and reads the diagnosis back to you with Watson Text-to-Speech. Any question outside a "find" request is simply answered by Granite (watsonx.ai) — so the same voice loop doubles as a general coding assistant. Safe by Design : Investigating agents are strictly read-only. Tests are never weakened or deleted. Every finding needs proof before it's reported. Fixes are applied one at a time, with full rollback if a fix doesn't turn the suite green. Built and Proven : Pure Python 3.8+, zero external dependencies Installable in one command (pip install .) — works in terminal, CI, VS Code, and PyCharm Validated across 15 real-world libraries: pandas, numpy, matplotlib, scikit-learn, requests, json, math, openpyxl, bs4, sqlalchemy, pydantic, pillow, tqdm, scipy, seaborn 14 paired demo files (bug + fix) plus dedicated test runner scripts From a red traceback to a spoken fix — in one Bob skill. GitHub : https://github.com/TheLoneSynapse/Bug-Whisperer
27 Sep 2026
