Sushar Hembram UNKN0WN006

`00 // identity`

I work on forecasting, quantitative research, machine learning and security systems — especially problems where the data is noisy, the environment changes, and confidence matters as much as the answer.

Most of that can be reduced to one question:

`P(outcome | evidence)`

Given the evidence available now, what should I believe — and what would make me change my mind?

`01 // current research`

_{The board separates work I am actively doing from research I am building toward. I would rather show a small honest pipeline than label every interest as “in progress.”}

`02 // selected work`

Maayavapi

threat intelligence / attacker behaviour

AI-assisted SSH honeypot and threat-intelligence system for turning hostile sessions into structured signals.

Built around

command and session telemetry
attack classification
IP reputation enrichment
geographic threat mapping
behavioural analysis
rule-based alerting

Python React TypeScript Supabase

NOESIS

repository security intelligence

Nested Orchestration of Exploitability & Structure Insight System

Repository-analysis system that reasons across:

authentication
authorization
dependencies
trust boundaries
data flow
architectural exploitability

Python FastAPI GitHub API Next.js

AletheiaIR

computer vision / meta-learning

self-correcting, contradiction-aware DFIR agent for Protocol SIFT

Protocol SIFT demonstrates what AI-assisted incident response looks like when it works. AletheiaIR addresses the next problem: what happens when the agent is confidently wrong?

AletheiaIR extends the Protocol SIFT agent loop with four additional pipeline stages that run before anything reaches the report:

Evidence Validation — every finding is checked against other source types. Memory and log both showing the same process chain raises confidence. A single source keeps it at "inferred."
Contradiction Detection — a deterministic rule engine checks for logical conflicts: a process visible in memory but absent from disk prefetch, C2 traffic with no owning process, claimed persistence without a registry artefact. Conflicts are recorded, not silenced.
Bounded Self-Correction — when contradictions exist, the agent reruns targeted analysis on the conflicting artefact types. Hard-capped at two iterations to prevent runaway loops.
Structured Audit Trail — every agent action is logged as JSONL with timestamps, confidence deltas, and tool call records. Judges can trace any finding back to its origin.

sha-256 protocol-rift

`03 // forecast lab`

I treat forecasting as a loop, not a one-shot prediction:

_{The values in the panel above are illustrative; the workflow is the important part. Public forecasting work is tracked separately from confidential platforms.}

`04 // security`

My security loop

recon → attack surface → vulnerability → exploitability → threat intelligence → response

Platforms / practice

OWASP OSINT Linux Penetration Testing

`05 // proof of work`

Signal	Evidence
COMSYS Hackathon IV	3rd Place · few-shot multimedia classification
Google SecOps MCP Challenge	Top 3 · security automation
NASA citizen science	2,800+ astronomical classifications
HackerOne + Intigriti	vulnerability research / security challenges
Google Cybersecurity Professional Certificate	completed professional certificate
Eureka! Junior — IIT Bombay	semifinalist

`06 // toolchain`

languages

quant / data / ML

systems / engineering

`07 // orbital activity`

The current year mapped onto a solar orbit: one marker = one calendar day.

Earth marks today and displays the current DAY n / 365 (or / 366 in a leap year).
Coloured markers are days with GitHub contributions; stronger activity produces brighter markers.
Future days remain dim until the year reaches them.
The lower panel shows recent public GitHub activity so the orbit is tied to actual work, not just decorative squares.
When the SVG is opened directly, contribution markers include date/count tooltips.

_{Generated daily from GitHub contribution data. Earth marks today; the orbit itself is the year map.}

`08 // developer signal`

A compact snapshot of my public GitHub activity: contribution intensity, commits, pull requests, issues, followers and repository footprint.

`09 // research platforms`

quantitative research & forecasting

security research & CTFs

competitive programming & problem solving

data science & machine learning

`10 // long game`

                     uncertainty
                         │
        ┌────────────────┼────────────────┐
        │                │                │
    forecasting       security         science
        │                │                │
      markets          systems        aerospace
        │                │                │
        └────────────────┼────────────────┘
                         │
                 decision systems

How do we understand, predict and engineer complex systems when the world refuses to stay still?

`11 // outside the terminal`

quant research · cybersecurity · aeronautics · science fiction · coffee · strange systems worth building

Navigation Menu