Sid
Sid Bakkireddy
besidd · open to work
sid@besidd — zsh — 80×24
sid@besidd:~$ /

now

Final semester at IIT (ISM) Dhanbad. Applying for AI/ML engineering roles — looking for a serious team building something real. Building in the meantime.

projects

.app
50+ users·solo full-stack·may 2025
grind-irl.app
Gamified productivity platform that turns daily habits into RPG quests with XP, streaks, and levels — founded and shipped end-to-end, bootstrapped to 50+ active users on a Firebase/GCP backend. grindirl.netlify.app ↗ · github ↗
.paper
~25% faster·patent-pending·IIT (ISM) Dhanbad
quantum-rl-ev.paper
Hybrid quantum-classical RL engine for EV charging allocation. Formulated routing as a QUBO, used Graph Attention Networks to compress the urban grid into quantum-compatible embeddings, then ran a hybrid QAOA + Grover-style filter to cut fleet wait time ~25%. Final-year research under Prof. Sukanta Halder. full paper ↗
.research
$300 · 1st place·Apart Research, London·jun 2024
llm-deception-detection.research
Catches GPT-3.5 sandbagging via response-time and token-usage metadata — deceptive outputs use ~25% fewer tokens. Won the Apart Research global hackathon onsite in London, which led to a 6-month AI safety fellowship. LessWrong ↗ · Alignment Forum ↗ · github ↗
.py
3-agent crew·CrewAI + Tavily
ai-usecase-generator.py
A CrewAI multi-agent crew — Researcher, Analyst, Collector — that turns one industry name into a usable set of applied GenAI use cases, using Tavily for retrieval and Streamlit for the surface. github ↗
.ipynb
8M reviews·71% precision·200k+ labeled
yelp-sentiment-bi.ipynb
Sentiment + business-intelligence pipeline over the full 8M Yelp review dataset. VADER baseline at 71% precision; XGBoost and logistic regression over 200k+ labeled points surfaced the features that actually correlate with a business doing well over time. github ↗
.ext
published·Chrome Web Store
brainstellar-tracker.ext
A published Chrome extension that tracks quant-interview logic puzzle progress on Brainstellar. Built it for myself while grinding the site, then shipped it publicly. Took an evening, saved many. Chrome Web Store ↗

more on github ↗.

writing

the log

2024-06hackathon.winWon the Apart Research ↗ Global AGI Deception Detection Hackathon onsite in London ($300 prize) — led to a 6-month AI safety fellowship working alongside researchers in the field.
2024-06paper.publishCo-authored Finding Deception in Language Models ↗ (also on Alignment Forum ↗).
2025-05project.shipFounded Grind IRL ↗ and bootstrapped it to 50+ active users solo.
2026-01hackathon.judgeSelected as judge for Apart Research's AI Manipulation Hackathon.
researchpatent.draftAuthored a patent draft + research report — Graph-Based Optimization and Learning for EV Charging Station Allocation ↗.
rankcompetitive.cpCodeforces Expert (1642) · LeetCode Knight (1875) · global rank 675 in Codeforces Round 185 (Div. 2).
chromeext.publishPublished the Brainstellar Progress Tracker ↗ on the Chrome Web Store.
lifehack.irlManaged the rental of my family's property remotely from my dorm — streamlined the paperwork, automated prospect filtering, only showed up to sign.
admitjee.advancedCleared JEE Advanced in the top 1% of ~1M applicants to secure a seat at IIT (ISM) Dhanbad.
sportphysicalState-level swimmer · football captain · consistent podium finisher in 5K races.

work with me

If you work at a big tech company or a crazy startup — I need you to refer me. Not ask. Not hope. Refer me.

I'm going to be one of the best AI engineers in the world. The log above is real. If your team needs someone who ships serious things and thinks carefully about AI: there's no reason not to. So just do it. Refer me. Help me out.

best way to reach me is a dm on x ↗ or email ↗.